[Federal Register Volume 71, Number 166 (Monday, August 28, 2006)]
[Rules and Regulations]
[Pages 50998-51048]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 06-7094]



[[Page 50997]]

-----------------------------------------------------------------------

Part II





Department of Transportation





-----------------------------------------------------------------------



National Highway Traffic Safety Administration



-----------------------------------------------------------------------



49 CFR Part 563



Event Data Recorders; Final Rule

Federal Register / Vol. 71, No. 166 / Monday, August 28, 2006 / Rules 
and Regulations

[[Page 50998]]


-----------------------------------------------------------------------

DEPARTMENT OF TRANSPORTATION

National Highway Traffic Safety Administration

49 CFR Part 563

[Docket No. NHTSA-2006-25666]
RIN 2127-AI72


Event Data Recorders

AGENCY: National Highway Traffic Safety Administration (NHTSA), 
Department of Transportation (DOT).

ACTION: Final rule.

-----------------------------------------------------------------------

SUMMARY: This final rule specifies uniform requirements for the 
accuracy, collection, storage, survivability, and retrievability of 
onboard motor vehicle crash event data in passenger cars and other 
light vehicles equipped with event data recorders (EDRs). This final 
rule responds to the growing practice in the motor vehicle industry of 
voluntarily installing EDRs in an increasing number of light vehicles. 
This final rule is intended to standardize the data obtained through 
EDRs so that such data may be put to the most effective future use and 
to ensure that EDR infrastructure develops in such a way as to speed 
medical assistance through providing a foundation for automatic crash 
notification (ACN). This final regulation: requires that the EDRs 
installed in light vehicles record a minimum set of specified data 
elements; standardizes the format in which those data are recorded; 
helps to ensure the crash survivability of an EDR and its data by 
requiring that the EDR function during and after the front and side 
vehicle crash tests specified in two Federal motor vehicle safety 
standards; and requires vehicle manufacturers to ensure the commercial 
availability of the tools necessary to enable crash investigators to 
retrieve data from the EDR. In addition, to ensure public awareness of 
EDRs, the regulation also requires vehicle manufacturers to include a 
standardized statement in the owner's manual indicating that the 
vehicle is equipped with an EDR and describing the functions and 
capabilities of EDRs.
    This final rule for standardization of EDR data will ensure that 
EDRs record, in a readily usable manner, the data necessary for ACN, 
effective crash investigations, and analysis of safety equipment 
performance. Standardization of EDR data will facilitate development of 
ACN, e-911, and similar systems, which could lead to future safety 
enhancements. In addition, analysis of EDR data can contribute to safer 
vehicle designs and a better understanding of the circumstances and 
causation of crashes and injuries.

DATES: Effective Date: This rule is effective October 27, 2006. The 
incorporation by reference of a certain publication listed in the 
regulation is approved by the Director of the Federal Register as of 
October 27, 2006.
    Compliance Dates: Except as provided below, light vehicles 
manufactured on or after September 1, 2010 that are equipped with an 
EDR and manufacturers of those vehicles must comply with this rule. 
However, vehicles that are manufactured in two or more stages or that 
are altered are not required to comply with the rule until September 1, 
2011.
    Petitions: If you wish to submit a petition for reconsideration of 
this rule, your petition must be received by October 12, 2006.

ADDRESSES: Petitions for reconsideration should refer to the docket 
number above and be submitted to: Administrator, Room 5220, National 
Highway Traffic Safety Administration, 400 Seventh Street, SW., 
Washington, DC 20590.

FOR FURTHER INFORMATION CONTACT: The following persons at the National 
Highway Traffic Safety Administration, 400 Seventh Street, SW., 
Washington, DC 20590.
    For technical and policy issues: Ms. Lori Summers, Office of 
Crashworthiness Standards (Telephone: 202-366-1740) (Fax: 202-493-
2739).
    For legal issues: Mr. Eric Stas, Office of the Chief Counsel 
(Telephone: 202-366-2992) (Fax: 202-366-3820).

SUPPLEMENTARY INFORMATION:

Table of Contents

I. Executive Summary
    A. Purpose of the Regulation
    B. Developments Culminating in the Notice of Proposed Rulemaking
    1. Early Agency Efforts on EDRs
    2. The Notice of Proposed Rulemaking
    C. Requirements of the Final Rule
    D. Lead Time
    E. Differences Between the Final Rule and the Notice of Proposed 
Rulemaking
    F. Impacts of the Final Rule
II. Background
    A. Overview of EDR Technology
    B. Chronology of Events Relating to NHTSA's Consideration of 
EDRs
    C. Petitions for Rulemaking
    1. Petitions From Mr. Price T. Bingham and Ms. Marie E. Birnbaum
    2. Petition From Dr. Ricardo Martinez
    D. October 2002 Request for Comments
III. Notice of Proposed Rulemaking
    A. Summary of the NPRM
    B. Summary of Public Comments to the NPRM
IV. The Final Rule and Response to Public Comments
    A. The Final Rule
    1. Summary of the Requirements
    2. Lead Time
    B. Response to Public Comments
    1. Whether NHTSA Should Require EDRs
    2. EDR Data Elements
    a. Number and Types of Required Data Elements
    b. The ``Acceleration'' and ``Delta-V'' Data Elements
    c. Multiple-event Crashes and the ``Multiple-event'' Data 
Element
    d. Sampling Rates and Recording Intervals for Required Data 
Elements
    3. EDR Data Standardization (Format) Requirements
    4. EDR Data Retrieval and Whether to Require a Standardized Data 
Retrieval Tool/Universal Interface
    5. EDR Survivability and Crash Test Performance Requirements
    6. Compliance Date
    7. Privacy Issues
    8. Owner's Manual Disclosure Statement
    9. Preemption
    10. Applicability of the EDR Rule to Multi-stage Vehicles
    11. Applicability of the EDR Rule to Heavy Vehicles and Buses
    12. Automatic Crash Notification and E-911
    13. Definitions
    a. ``Trigger Threshold''
    b. ``Event''
    c. ``Event Data Recorder''
    14. Utilization of SAE and IEEE Standards
    15. Costs
    16. Other Issues
    a. Scope and Purpose
    b. Technical Changes to Definitions and New Definitions
    c. Data Capture
    d. Miscellaneous Comments
V. Rulemaking Analyses and Notices

I. Executive Summary

A. Purpose of the Regulation

    Event data recorders have been used in recent years in a variety of 
transportation modes to collect crash information. EDR data will play 
an increasing role in advancing developing networks for providing 
emergency medical services. Specifically, EDR data can help the safety 
community develop ACN, electronic 911 (e-911), and other emergency 
response systems to improve medical services to crash victims. In 
addition, EDR data can also provide information to enhance our 
understanding of crash events and safety system performance, thereby 
potentially contributing to safer vehicle designs and more effective 
safety regulations.
    EDRs have experienced dramatic changes in the past decade, both in 
terms of their technical capabilities and fleet penetration. EDRs today 
demonstrate a range of features, with some systems collecting only 
vehicle acceleration/deceleration data, but

[[Page 50999]]

others collecting these data plus a host of complementary data such as 
driver inputs (e.g., braking and steering) and vehicle system status. 
The challenge for NHTSA has been to devise an approach that would 
encourage broad application of EDR technologies in motor vehicles and 
maximize the usefulness of EDR data for the medical community, 
researchers, and regulators, without imposing unnecessary burdens or 
hampering future improvements to EDRs.
    In light of the relatively high new vehicle fleet penetration of 
EDRs (currently estimated at 64%) and present trends, we do not believe 
that it is necessary to mandate the installation of EDRs in all new 
vehicles. Were these trends reversed or slowed, we would consider 
revisiting this assessment. For now, we believe that standardization of 
EDR data represents the most important area of opportunity in terms of 
enhancing the yield of benefits from EDRs. We recognize that the 
automobile industry has already invested considerable effort and 
resources into developing effective EDR technologies, so we want to be 
especially careful not to adopt requirements that would result in 
unnecessary costs.
    Accordingly, this final rule regulates voluntarily-provided EDRs by 
specifying a minimum core set of required data elements and 
accompanying range, accuracy, and resolution requirements for those 
elements. This will help ensure that EDRs provide the types of data 
most useful for the emergency medical services (EMS) community and 
crash reconstructionists, and in a manner that promotes the consistency 
and comparability of these data. We note that by specifying this 
minimum data set, we are not limiting manufacturers' ability to design 
EDRs that collect a broader set of data, provided that the required 
elements are present.
    The rule also includes requirements for the survivability of EDR 
data (so that it is not lost in most crashes) and the retrievability of 
EDR data (so that it can be obtained by authorized users). In sum, the 
objectives of our regulation are to get the right data, in sufficient 
quantity and in a standardized format, and to ensure that the data can 
survive most crash events and be retrieved by intended users.
    By promulgating a uniform national regulation for EDRs, it is our 
intent to provide one consistent set of minimum requirements for 
vehicle manufacturers that choose to install EDRs. We believe that this 
approach will not only enhance the quality of EDR data, but also 
facilitate increased numbers of new light vehicles equipped with EDRs. 
We also believe that this minimum data set provides key elements in a 
standardized format that will be useful for ACN or other telematic 
systems.

B. Developments Culminating in the Notice of Proposed Rulemaking

1. Early Agency Efforts on EDRs
    NHTSA has been assessing the potential benefits of EDR for over a 
decade, and in that time, we have witnessed a significant maturation of 
EDR technology. The agency initially began examining EDRs in 1991 as 
part of the Special Crash Investigations (SCI) program. In 1997, the 
National Transportation Safety Board (NTSB) and the National 
Aeronautics and Space Administration (NASA) Jet Propulsion Laboratory 
(JPL) recommended that NHTSA consider the possibility of requiring the 
installation of EDRs in motor vehicles. NTSB made additional 
recommendations related to EDRs in 1999 (i.e., suggesting that EDRs be 
installed in school buses and motor coaches). Since 1998, NHTSA has 
sponsored two Working Groups to examine and report on EDR issues.
    As discussed below, the agency received two petitions for 
rulemaking in the late 1990s asking that light vehicles be equipped 
with ``black boxes'' (i.e., EDRs) that would record data during a crash 
so that it could be read later by crash investigators. However, the 
agency denied those petitions because the industry was already moving 
voluntarily in the direction recommended by the petitioners, and 
because the agency believed that certain outstanding issues would best 
be addressed in a non-regulatory context.
    In 2001, NHTSA received a third petition for rulemaking related to 
EDRs from Dr. Ricardo Martinez, seeking a requirement for installation 
of EDRs as well as standardization of EDR data. After considering the 
Martinez petition and the current situation vis-[agrave]-vis EDRs, we 
decided to publish a request for comments as to what future role the 
agency should take related to the continued development and 
installation of EDRs in motor vehicles. This notice was published on 
October 11, 2002 (67 FR 63493), and after considering the input from a 
variety of interested stakeholders and the public, we decided to grant 
the Martinez petition in part (i.e., the request for standardization 
and retrievability) and to deny it in part (i.e., the request for an 
EDR mandate).
2. The Notice of Proposed Rulemaking
    On June 14, 2004, NHTSA published a notice of proposed rulemaking 
(NPRM) proposing requirements for EDRs voluntarily installed by light 
vehicle manufacturers (69 FR 32932).\1\ The decision to conduct 
rulemaking reflected careful deliberation and our belief that EDRs 
represent a significant technological safety innovation, particularly 
for the emergency response safety community.\2\ Again, the proposal 
sought to standardize the elements and format of data deemed most 
appropriate for advancing our goals of enabling ACN and improving crash 
reconstructions and for ensuring the retrievability of that 
information. Most of these data elements are already recorded by 
current EDRs. It was not our intention to require an exhaustive list of 
non-essential data elements that would significantly increase the cost 
of EDRs, thereby jeopardizing the current, high rate of installation.
---------------------------------------------------------------------------

    \1\ Docket No. NHTSA-2004-18029-2.
    \2\ We note that NHTSA has been assessing the potential benefits 
of EDRs for over a decade, and in that time, we have witnessed a 
significant maturation of EDR technology. For further information on 
these agency research and analytical efforts, please consult the 
NPRM, which discussed this topic extensively (see 69 FR 32932, 32933 
(June 14, 2004)).
---------------------------------------------------------------------------

    In summary, the NPRM proposed to require light vehicles voluntarily 
equipped with an EDR to meet uniform, national requirements for the 
collection, storage, and retrievability of onboard motor vehicle crash 
event data. The proposal included Table I, Data Elements Required for 
All Vehicles Equipped with an EDR, which included 18 required elements 
that would have to be recorded during the interval/time and at the 
sample rate specified in that table. The proposal also included Table 
II, Data Elements Required for Vehicles Under Specified Conditions, 
which included 24 elements that would have to be recorded (during the 
interval/time and at the sample rate specified in that table) if the 
vehicle is equipped with certain devices or is equipped to measure 
certain elements. Table III, Recorded Data Element Format, included 
proposed range, accuracy, precision, and filter class requirements for 
each data element.
    The NPRM also proposed a methodology for data capture under 
specified conditions and circumstances (i.e., providing a hierarchy for 
when new EDR data would overwrite existing data already stored in 
memory). Simply put, EDRs are constantly monitoring a variety of 
vehicle systems and parameters when the vehicle is in operation, but 
the devices only have a limited amount of short-term (volatile) memory 
and long-term (non-volatile) memory available for recording for these

[[Page 51000]]

purposes. So when vehicle manufacturers develop EDRs, they must make 
judgments as to which data are the most important to be captured and 
recorded (e.g., events surrounding the deployment of an air bag are 
generally regarded as very important). Frequently, data stored in non-
volatile memory are over-written (replaced) or deleted. The NPRM's 
proposed provisions related to data capture were intended to ensure 
that EDRs not only capture data according to a uniform methodology, but 
also that the methodology maximizes the generation of data suitable for 
the agency's safety purposes.
    Because data standardization is only beneficial if the data can be 
retrieved and used, the agency decided to address the issue of data 
retrievability as part of our rulemaking. The NPRM also proposed to 
require vehicle manufacturers to submit sufficient non-proprietary 
technical information to the public docket as would permit third 
parties to manufacture a device capable of accessing, interpreting, and 
converting the data stored in the EDR. Under the proposal, such 
information would be required to be submitted to the docket not later 
than 90 days prior to the start of production of the EDR-equipped 
vehicle makes and models to which the information relates, and vehicle 
manufacturers would be required to keep that information updated, by 
providing information not later than 90 days prior to making any 
changes that would make the previously submitted information no longer 
valid. However, as discussed in the NPRM, our proposal offered one 
possible way to handle the data retrievability issue, and we sought 
comment on alternative approaches.
    In addition, the NPRM proposed survivability requirements for EDR 
data when the vehicle is crash tested under existing testing 
requirements of Federal Motor Vehicle Safety Standard (FMVSS) Nos. 208, 
Occupant Crash Protection, 214, Side Impact Protection, and 301, Fuel 
System Integrity, and it also proposed to require that the data be 
retrievable by the methodology specified by the vehicle manufacturer 
for not less than 30 days after the test and without external power.
    Finally, the NPRM proposed a specific owner's manual statement 
related to EDRs that would make members of the public aware when their 
vehicle is equipped with an EDR and also explain the intended purpose 
of the EDR and how it operates.

C. Requirements of the Final Rule

    After careful consideration of the public comments on the NPRM, we 
are promulgating this final rule to establish a regulation for 
voluntarily-installed EDRs in order to standardize EDR data. The 
approach of this final rule is generally consistent with that of the 
NPRM, although we have further tailored the requirements of the 
regulation to advance the stated purposes of this rulemaking without 
requiring substantial costs or impeding the technological development 
of EDRs. We believe that with certain modest modifications, many 
current EDR systems can meet our goals of facilitating ACN and 
improving crash reconstructions.
    In overview, the final rule specifies uniform, national 
requirements for light vehicles voluntarily equipped with EDRs, 
including the collection, storage, and retrievability of onboard motor 
vehicle crash event data. It also specifies requirements for vehicle 
manufacturers to make tools and/or methods commercially available so 
that authorized crash investigators and researchers are able to 
retrieve data from such EDRs.
    Specifically, the regulation applies to passenger cars, 
multipurpose passenger vehicles, trucks, and buses with a gross vehicle 
weight rating (GVWR) of 3,855 kg (8,500 pounds) or less and an unloaded 
vehicle weight of 2,495 kg (5,500 pounds) or less, except for walk-in 
van-type trucks or vehicles designed to be sold exclusively to the U.S. 
Postal Service, that are equipped with an event data recorder and to 
manufacturers of these vehicles. Subject to an exception for final-
stage manufacturers and alterers discussed below, compliance with the 
requirements of the final rule commences for covered vehicles 
manufactured on or after September 1, 2010. The final rule is intended 
to be technology-neutral, so as to permit compliance with any available 
EDR technology that meets the specified performance requirements.
    The following points highlight the key provisions of the final 
rule:
     Each vehicle equipped with an EDR must record all of the 
data elements listed in Table I, during the interval/time and at the 
sample rate specified in that table. There are 15 required data 
elements (see paragraph 563.7(a), Table I). Examples of these data 
elements are ``delta-V, longitudinal,'' ``maximum delta-V, 
longitudinal,'' ``speed, vehicle indicated,'' and ``safety belt status, 
driver.''
     Each vehicle equipped with an EDR that records any of the 
data elements listed in Table II identified as if recorded (most 
elements in that table) must capture and record that information 
according to the interval/time and at the sample rate specified in that 
table. Data elements listed in Table II as ``if equipped'' (i.e., 
``frontal air bag deployment, time to nth stage, driver'' 
and ``frontal air bag deployment, time to nth stage, right 
front passenger'') must record the specified information if they are 
equipped with the relevant item, even if they are not presently doing 
so.\3\ There are 30 data elements included in Table II (see paragraph 
563.7(b), Table II). Examples of these data elements are ``lateral 
acceleration,'' ``longitudinal acceleration,'' ``frontal air bag 
suppression switch status, right front passenger (on, off, or auto), 
and safety belt status, right front passenger (buckled, not buckled).
---------------------------------------------------------------------------

    \3\ The ``frontal air bag deployment, time to n\th\ stage'' data 
elements provide critical timing data for vehicles equipped with 
multi-stage air bags, which will help in assessing whether an air 
bag is deploying correctly during a crash (i.e., whether the sensors 
are functioning properly). In drafting this final rule, we had 
considered including these two elements as required elements under 
Table I, but we recognized that not all vehicles are equipped with 
multi-stage air bags. Thus, by including these elements in Table II 
and requiring recording of that information if the vehicle is so 
equipped, we are, in effect, requiring this data from all vehicles 
equipped with an EDR and multi-stage air bags.
---------------------------------------------------------------------------

     The data elements required to be collected by the EDR 
pursuant to Tables I and II, as applicable, must be recorded in 
accordance with the range, accuracy, and resolution requirements 
specified in Table III, Recorded Data Element Format (see paragraph 
563.8(a), Table III).
     For EDRs that record acceleration, the longitudinal and 
lateral acceleration time-history data must be filtered in accordance 
with the filter class specified in Table III (i.e., Society of 
Automotive Engineers (SAE) Recommended Practice J211-1, March 1995, 
``Instrumentation For Impact Test--Part 1--Electronic Instrumentation'' 
(SAE J211-1, Class 60), which the regulation incorporates by reference 
(see paragraph 563.8(b)). Such filtering may be done during collection 
or post-processing.
     The EDR must collect and store data elements for events in 
accordance with the following conditions and circumstances as specified 
in paragraph 563.9:
    (1) In an air bag deployment crash, the data recorded from any 
previous crash must be deleted; the data related to the deployment must 
be recorded, and the memory must be locked in order to prevent any 
future overwriting of these data.

[[Page 51001]]

    (2) In an air bag non-deployment crash that meets the trigger 
threshold, all previously recorded data in the EDR's memory must be 
deleted from the EDR's memory, and the current data (up to two events) 
must be recorded.
     In order to ensure the survivability of EDR data in most 
crashes, the EDR is tested in conjunction with crash tests already 
required under FMVSS No. 208, Occupant Crash Protection, and FMVSS No. 
214, Side Impact Protection (see paragraph 563.10). Except for elements 
discussed below, the data elements required under Tables I and II must 
be recorded in a specified format, must exist at the completion of the 
crash test, and must be retrievable by a methodology specified by the 
vehicle manufacturer for not less than 10 days after the test.
    The EDR is not required to meet the above survivability 
requirements for the following data elements: (1) ``Engine throttle, % 
full,'' (2) ``service brake, on/off,'' and (3) ``engine RPM.'' These 
elements have been excluded from these requirements because vehicles 
are crash tested without the engine running for safety reasons, so the 
EDR would not be able to record the above data elements under those 
circumstances.
     For vehicles equipped with an EDR, vehicle manufacturers 
must include a specified statement in the owner's manual to make the 
operator aware of the presence, function, and capabilities of the EDR.
     In order to ensure the retrievability of EDR data, each 
vehicle manufacturer that installs EDRs must ensure by licensing 
agreement or other means that the necessary tool(s) are commercially 
available for downloading the required EDR data. The tool must be 
commercially available not later than 90 days after the first sale of 
the vehicle for purposes other than resale.

D. Lead Time

    In order to limit the transition costs associated with the 
standardization of EDR data, we sought in the NPRM to provide adequate 
lead time to manufacturers to enable them to incorporate necessary 
changes as part of their routine production cycles. To that end, the 
NPRM proposed a compliance date of September 1, 2008 for the EDR 
regulation. However, vehicle manufacturers commented that the lead time 
in the proposed rule would be inadequate to allow manufacturers to 
incorporate the necessary changes as part of their regular production 
cycle. Those commenters argued that a longer lead time is needed to 
minimize the costs and burdens associated with the EDR rule, 
particularly for those manufacturers which have already incorporated 
EDRs in a large proportion of their fleets.
    After carefully considering the public comments on lead time, we 
have decided to require covered vehicles manufactured on or after 
September 1, 2010 to comply with the requirements of this final rule, 
subject to the exception below. Again, it is our intention to limit the 
costs associated with this final rule for the standardization of EDR 
data, including implications associated with new definitions, new pre-
crash data collection, data download strategies, and data element costs 
associated with meeting the range and accuracy requirements. We believe 
that a lead time in excess of four years should prove adequate for all 
vehicle manufacturers and all vehicle lines, without the need for a 
phase-in. Vehicle manufacturers may voluntarily comply with these 
requirements prior to this date.
    Consistent with the policy set forth in NHTSA's February 14, 2005 
final rule on certification requirements under Federal motor vehicle 
safety standards for vehicles built in two or more stages and altered 
vehicles (70 FR 7414), we are providing final-stage manufacturers and 
alterers that produce vehicles covered by this regulation with an extra 
year to comply. Accordingly, these manufacturers must meet the 
requirements of this final rule for vehicles manufactured on or after 
September 1, 2011. However, final-stage manufacturers and alterers may 
voluntarily comply with the requirements of the regulation prior to 
this date.

E. Differences Between the Final Rule and the Notice of Proposed 
Rulemaking

    As noted above, NHTSA has decided to issue the present final rule 
to standardize EDR data in order to further our stated purposes of 
ensuring that EDRs record the data necessary for effective 
implementation of ACN, crash investigations, and analysis of safety 
equipment performance. In order to achieve these objectives (and to 
garner the derivative benefits that EDR-generated data may provide in 
terms of safer vehicle designs), we have largely retained the general 
approach presented in the NPRM. However, after further study and a 
careful review of the public comments, we have decided to make a number 
of modifications as part of the final rule in order to better reflect 
the current state of EDR technology and the data elements (including 
form and format) that will meet our research and policy objectives in a 
manner that is both effective and practicable.
    The main differences between the NPRM and the final rule involve a 
change in the definition of ``event data recorder,'' selection of data 
elements (i.e., which elements are required), changes to the range/
accuracy/resolution requirements, modification of the test requirements 
related to EDR survivability, and extension of lead time for 
implementing the regulation. A number of minor technical modifications 
are also incorporated in the final rule in response to public comments 
on the NPRM. All of these changes and their rationale are discussed 
fully in the balance of this document. However, the following points 
briefly describe the main differences between the NPRM and this final 
rule.
     In the NPRM, the term ``event data recorder'' was defined 
as ``a device or function in a vehicle that records any vehicle or 
occupant-based data just prior to or during a crash, such that the data 
can be retrieved after the crash. For purposes of this definition, 
vehicle or occupant-based data include any of the data elements listed 
in Table I of this part.'' However, several commenters stated that 
under this definition, virtually all vehicles would be considered to 
have an EDR, because most vehicles capture freeze-frame data required 
for internal processing; therefore, commenters argued that the proposed 
definition is overly broad (i.e., covering vehicles not equipped with a 
true EDR) and would create a de facto mandate for EDRs, contrary to the 
agency's expressed intent. Therefore, in this final rule, we have 
revised the definition of ``event data recorder'' to read as follows: 
``a device or function in a vehicle that records the vehicle's dynamic, 
time-series data during the time period just prior to a crash event 
(e.g., vehicle speed vs. time) or during a crash event (e.g., delta-V 
vs. time), intended for retrieval after the crash event. For the 
purposes of this definition, the event data do not include audio and 
video data.''
     In the final rule, we have decided to make certain 
modifications to the proposed tables of EDR data elements. Table I, 
Data Elements Required For All Vehicles Equipped With an EDR, has been 
amended by deleting five data elements (i.e., (1) longitudinal 
acceleration (moved to Table II); (2) engine RPM (moved to Table II); 
(3) frontal air bag deployment level, driver; (4) frontal air bag 
deployment level, right front passenger, and (5) time from event 2 to 
3) and by adding two data

[[Page 51002]]

elements (i.e., (1) time, maximum delta-V, and (2) delta-V, 
longitudinal).
    Table II, Data Elements Required for Vehicles under Specified 
Conditions, has been modified in two ways from the NPRM. First, the 
data elements now listed in Table II as ``if recorded'' will be 
required only if the data elements are recorded by the EDR (i.e., 
stored in non-volatile memory as would permit later retrieval), rather 
than the NPRM's approach which would have required those elements if 
the vehicle were equipped to measure those elements. However, for the 
final rule's data elements listed in Table II as ``if equipped,'' a 
manufacturer's EDRs must record the specified information, even if its 
current EDRs are not doing so.
    Furthermore, Table II has been amended by adding six data elements 
(i.e., Table II includes four new elements: (1) Lateral delta-V; (2) 
lateral cumulative maximum delta-V; (3) time to cumulative maximum 
lateral delta-V, and (4) time to cumulative maximum resultant delta-V. 
In addition, as indicated above, two items have been moved from Table I 
to Table II: (1) Longitudinal acceleration; and (2) engine RPM.).
     In the NPRM, we proposed a definition for ``trigger 
threshold,'' the point at which a recordable event is recognized by the 
EDR, as a ``change in vehicle velocity * * * that equals or exceeds 0.8 
km/h within a 20 ms interval.'' That definition encompassed movement in 
either a longitudinal or lateral direction.
    In the final rule, we decided to change the definition of ``trigger 
threshold'' for the longitudinal direction to ``a change in vehicle 
velocity * * * that equals or exceeds 8 km/h within a 150 ms 
interval.'' For vehicles whose EDRs measure lateral delta-V or lateral 
acceleration, we are using the same trigger threshold. In the final 
rule, we have changed the definition of ``time zero'' to account for 
different EDR crash detection strategies (i.e., using a ``wake-up'' 
time for EDRs that wake up just as a crash starts, or a change in 
velocity over a short period for EDRs that are continuously running). 
We have also added a new definition for ``end of event time.'' ``Time 
zero'' and ``end of event time'' are defined in a manner consistent 
with SAE J1698.
     In the final rule, we have changed our approach in terms 
of the type of data that an EDR may capture to assess crash severity. 
Specifically, the NPRM proposed to require EDRs to measure vehicle 
acceleration, but the final rule requires the EDR to record delta-V. 
However, if the EDR records acceleration data in non-volatile memory, 
that information must also be captured and recorded under the final 
rule.
     As part of the final rule, the agency has decided to 
reduce the number of events that must be recorded in a multi-event 
crash from three (as proposed in the NPRM) to two.
     For each of the proposed data elements (when applicable), 
the NPRM specified a recording interval and sampling rate in order to 
standardize EDR data across the spectrum of new light vehicles. We have 
decreased the pre-crash recording interval from 8 seconds prior to the 
crash, as proposed in the NPRM, to 5 seconds prior to the crash, and we 
have reduced the amount of time allocated for collecting crash data 
from 0.5 second, as proposed in the NPRM, to 0.25 second in this final 
rule.
     The final rule has modified the NPRM s data format 
requirements, which proposed to require covered data elements to be 
recorded in accordance with the range, accuracy, precision, and filter 
class specified in Table III, Recorded Data Element Format, where 
applicable. The major changes were: (1) To reduce the maximum range for 
acceleration measurements from 100 G maximum, as proposed in the NPRM, 
to 50 G maximum, and (2) to reduce the required accuracy of these same 
devices (and the data generated therefrom) from within 1 
percent, as proposed in the NPRM, to within 5 percent.
     After requesting comments on alternate approaches in the 
NPRM, the agency has adopted a different approach for ensuring that 
manufacturers make sufficient information available to permit EDR data 
to be downloaded by potential users. The NPRM proposed to require 
vehicle manufacturers make publicly available sufficient information to 
permit third parties to build a retrieval tool for EDR data by 
submitting such materials to the NHTSA Docket (and keeping such 
information updated). However, in the final rule, we have decided, 
consistent with manufacturers' comments, to require manufacturers to 
ensure by licensing agreement or other means that retrieval tools for 
EDR data are commercially available.
     In the NPRM, we proposed to require manufacturers to send 
detailed information on an ongoing basis to the agency about retrieval 
tools for EDR data. However, in the final rule, we have decided to 
require vehicle manufacturers to ensure that EDR retrieval tools are 
commercially available, something which manufacturers may accomplish 
either by producing the tools themselves or working directly with their 
suppliers through licensing agreements. Accordingly, the need for 
reports to the agency, as contemplated in the NPRM, no longer exists.
     The final rule clarifies that EDR survivability testing 
will be conducted without the engine running, in order to prevent a 
potentially hazardous situation for testing personnel and facilities. 
The final rule specifies that the ``engine throttle,'' ``service brake, 
on/off,'' and (3) ``engine RPM'' data elements are not required to be 
recorded as part of survivability testing. While we are retaining the 
general approach for survivability testing, we are decreasing the 
number of tests required to demonstrate survivability. Under the NPRM, 
we proposed using FMVSS Nos. 208 (frontal), 214 (side), and 301 (rear) 
tests, but in the final rule, we have decided to delete the requirement 
for the Standard No. 301 test.
     We have decided as part of the final rule to extend the 
lead time for compliance by covered vehicles by two years, until 
September 1, 2010. In addition, the final rule sets the compliance date 
for final-stage manufacturers and alterers at one year beyond the 
compliance deadline for other manufacturers (i.e., September 1, 2011).

F. Impacts of the Final Rule

    It is difficult for the agency to quantify the benefits expected to 
result from this final rule for standardization of EDR data. That is 
because the EDR devices themselves are not designed to be systems for 
crash avoidance or crashworthiness, but instead they offer an important 
tool to enable better EMS response and to better understand crashes and 
crash-related events. However, it is possible to describe the benefits 
of EDRs in qualitative terms.
    To the extent that EDR data are compatible with developing ACN and 
e-911 systems, emergency medical personnel are likely to arrive at a 
crash site better informed and thus better prepared to deal with the 
injuries they encounter. Because expedient and appropriate post-crash 
medical care is often critical to achieving the best possible outcome 
for the injured person, we believe that EDR data have the potential to 
make a positive contribution in this area.
    We also believe that EDRs can provide important benefits by giving 
researchers a relatively inexpensive way of obtaining higher quality 
data and thus a more accurate and detailed understanding of the 
circumstances surrounding crashes, including how the vehicles and their 
safety systems performed. In many cases, such

[[Page 51003]]

information may be derived from crash reconstructions, but such 
measurements tend to be reasoned estimates, as compared to the directly 
measured data provided by the EDR. There is certain information, such 
as how the air bag deployed (e.g., low level or high level) or when it 
deployed, that cannot be determined without an EDR. To the extent that 
EDRs help researchers and policymakers to better understand the events 
surrounding crashes, NHTSA and vehicle manufacturers will be better 
able to develop effective safety countermeasures as reflected in 
Federal motor vehicle safety standards and new vehicle designs.
    In sum, we believe that having a uniform and standardized data set 
for EDRs will increase the compatibility, comparability, and overall 
usefulness of EDR data, which will benefit the public directly through 
the availability of ACN and e-911, and indirectly through improved 
crash information for research and regulatory efforts.
    In terms of costs, we believe that the costs of this final rule 
should be minimal, averaging up to $0.17 per vehicle. Several factors 
contribute to this result. First, we estimate that about 64 percent of 
new light vehicles in 2005 are already equipped with EDRs, which have 
been provided by adding the EDR capability to the vehicles' air bag 
control systems. Thus, EDRs largely capture information that is already 
being processed by the vehicle, so EDRs are not responsible for the 
much higher costs of sensing much of the data in the first place. 
Therefore, the costs of this final rule reflect the incremental costs 
for vehicles voluntarily equipped with EDRs to comply with the 
requirements of the regulation.
    Second, the agency has sought to limit the number of EDR data 
elements and associated requirements to the minimum necessary to 
achieve our stated purposes. We have determined that the industry's 
current state-of-the-art largely meets our purposes, so we have found 
it generally unnecessary to specify requirements for additional sensors 
or other hardware that would increase EDR costs appreciably. (The most 
significant technology cost may involve the need to upgrade EDR memory 
chips.) Furthermore, we expect that administrative costs and compliance 
costs will be negligible.
    In sum, for the 64 percent of new light vehicles already equipped 
with an EDR, the estimated total cost to comply with the requirements 
of this final rule (i.e., Table I data elements) will range up to $1.7 
million. If we were to assume that all 15.5 million new light vehicles 
were equipped with EDRs, the estimated total cost will range up to 
$10.9 million.

II. Background

A. Overview of EDR Technology

    Event data recorders capture vehicle crash information.\4\ Basic 
EDRs capture only vehicle acceleration/deceleration data, while more 
sophisticated EDRs capture these data plus a host of complementary 
data, such as driver inputs (e.g., braking and steering) and the status 
of vehicle safety systems (e.g., seatbelt pretensioners).
---------------------------------------------------------------------------

    \4\ The term ``EDR'' can be used to describe many different 
types of devices. For this final rule, the term EDR means a device 
or function in a vehicle that captures the vehicle's dynamic, time-
series data during the time period just prior to a crash event 
(e.g., vehicle speed vs. time) or during a crash event (e.g., delta-
V vs. time), such that the data can be retrieved after the crash 
event. For the purposes of this definition, the event data do not 
include audio and video data.
---------------------------------------------------------------------------

    The EDR captures crash data by monitoring several of the vehicle's 
systems, such as brakes, air bags, and seat belts. It continuously 
captures and overwrites (erases) information on these systems so that a 
record of the most recent period (up to a few seconds) is always 
available. If an ``event'' occurs (i.e., a crash meeting a pre-
determined threshold of severity), then the EDR moves captured pre-
crash information (up to a few seconds) into its long-term memory. EDRs 
also record (in long-term memory) data after the start of the crash (up 
to a few seconds), such as the timing and manner of the deployment of 
the air bags.
    EDRs have been installed as standard equipment in most light motor 
vehicles in recent years, particularly vehicles with air bags. We 
estimate that 64 percent of model year (MY) 2005 passenger cars and 
other light vehicles have some recording capability, and that more than 
half record data elements such as crash pulse data. This is based on 
manufacturer reports regarding their 2005 vehicles and then weighted 
using 2003 corporate-level vehicle sales figures to determine a fleet 
average.

B. Chronology of Events Relating to NHTSA's Consideration of EDRs

    In 1991, NHTSA's Special Crash Investigations program first 
utilized EDR information in an agency crash investigation. General 
Motors, the vehicle's manufacturer, cooperated with the program. 
Throughout the 1990s, NHTSA's SCI team utilized EDRs as one of their 
investigative tools, and from 1991 through 1997, SCI worked with 
manufacturers to read approximately 40 EDRs. Starting around 2000, the 
collection of EDR data was automated, and to date, NHTSA's crash 
investigation programs have collected information on about 2,700 
crashes with EDR files.
    The National Transportation Safety Board has also played a role in 
agency efforts related to event data recorders. The NTSB has been 
active in data recorders for a long time, first concentrating on 
aircraft and later on railroads and ships. More recently, NTSB has been 
active in the area of EDRs for highway vehicles. In 1997, the Safety 
Board issued its first highway vehicle EDR-related Safety 
Recommendation, H-97-18,\5\ to NHTSA, recommending that the agency 
``pursue crash information gathering using EDRs.'' NTSB recommended 
that the agency ``develop and implement, in conjunction with the 
domestic and international automobile manufacturers, a plan to gather 
better information on crash pulses and other crash parameters in actual 
crashes, utilizing current or augmented crash sensing and recording 
devices.'' NTSB subsequently closed this recommendation, citing NHTSA's 
actions as acceptable. Also in that year, the National Aeronautics and 
Space Administration (NASA) Jet Propulsion Laboratory (JPL), in a study 
conducted for NHTSA about advanced air bag technology, recommended that 
the agency ``study the feasibility of installing and obtaining crash 
data for safety analyses from crash recorders on vehicles.''
---------------------------------------------------------------------------

    \5\ NTSB public forum on air bags and child passenger safety 
(March 1997). See http://www.ntsb.gov/publictn/1997/rp9701.pdf.
---------------------------------------------------------------------------

    In early 1998, NHTSA's Office of Vehicle Safety Research formed an 
EDR Working Group comprised of members from industry, academia, and 
other government organizations. The working group was formed in 
response to NHTSA's growing interest in EDRs, the NTSB's 
recommendation, and interest from vehicle manufacturers. The group's 
objective was to facilitate the collection and utilization of collision 
avoidance and crashworthiness data from on-board EDRs. The NHTSA-
sponsored EDR Working Group published a final report on the results of 
its deliberations in August 2001.\6\ The working group found that EDRs 
have the potential to greatly improve highway safety, for example, by 
improving occupant protection systems and improving the accuracy of 
crash reconstruction.
---------------------------------------------------------------------------

    \6\ Event Data Recorders, Summary of Findings by the NHTSA EDR 
Working Group, August 2001, Final Report (Docket No. NHTSA-99-5218-
9).
---------------------------------------------------------------------------

    In 1999, NTSB issued a second set of recommendations to NHTSA 
related to

[[Page 51004]]

EDRs (H-99-53 and H-99-54 \7\) recommending that the agency require 
standardized EDRs to be installed on school buses and motor coaches. In 
2000, NHTSA responded to these NTSB recommendations by sponsoring a 
second working group related to EDRs--the NHTSA Truck & Bus EDR Working 
Group. This Working Group collected facts related to use of EDRs in 
trucks, school buses, and motor coaches--a natural follow-up activity 
from the first working group that concentrated on light vehicles. The 
final report of the NHTSA Truck and Bus EDR Working Group was published 
in May 2002.\8\
---------------------------------------------------------------------------

    \7\ Bus Crashworthiness Issues, Highway Special Investigation 
Report (NTSB/SIR-99/04) (Washington, DC (1999)). See http://www.ntsb.gov/publictn/1999/sir9904.pdf.
    \8\ Event Data Recorders, Summary of Findings by the NHTSA EDR 
Working Group, May 2002, Final Report, Volume II, Supplemental 
Findings for Trucks, Motor Coaches, and School Buses. (Docket No. 
NHTSA-2000-7699-6).
---------------------------------------------------------------------------

    In 2004, NTSB issued EDR recommendations to NHTSA for a third time. 
This set of recommendations was prompted by a crash that occurred at a 
farmers' market in Santa Monica, CA, which resulted in multiple deaths. 
In examining that crash, the Safety Board found that they could not 
determine exactly what occurred with respect to the driver controls and 
indicated that EDRs should be installed on all new vehicles. 
Recommendation H-04-26 \9\ reads: ``Once standards for event data 
recorders are developed, require their installation in all newly 
manufactured light-duty vehicles.'' In 2005, NHTSA sent a letter to the 
Safety Board asking them to reconsider their recommendation, indicating 
that many new cars and light trucks are already equipped with EDRs and 
that standardization of installed EDRs is the main issue, which is 
being addressed by this final rule.
---------------------------------------------------------------------------

    \9\ Rear-End Collision and Subsequent Vehicle Intrusion into 
Pedestrian Space at Certified Farmers' Market Santa Monica, 
California (July 16, 2003). See http://www.ntsb.gov/publictn/2004/har0404.pdf.
---------------------------------------------------------------------------

    For further information, NHTSA has developed a website about 
highway-based EDRs located at the following address: http://www-nrd.nhtsa.dot.gov/edr-site/index.html.

C. Petitions for Rulemaking

1. Petitions From Mr. Price T. Bingham and Ms. Marie E. Birnbaum
    In the late 1990s, the agency denied two petitions for rulemaking 
asking us to require the installation of EDRs in new motor vehicles 
(see 63 FR 60270 (November 9, 1998) and 64 FR 29616 (June 2, 1999)).
    The first petition, submitted by Mr. Price T. Bingham,\10\ a 
private individual, asked the agency to initiate rulemaking to require 
air bag sensors to record data during a crash so that it could later be 
read by crash investigators. The petitioner cited a concern about air 
bag deployments that might be ``spontaneous,'' but he did not limit the 
petition to that issue.
---------------------------------------------------------------------------

    \10\ Docket No. NHTSA-1998-4368-1.
---------------------------------------------------------------------------

    The second petition, submitted by Ms. Marie E. Birnbaum,\11\ also a 
private individual, asked us to initiate rulemaking to require 
passenger cars and light trucks to be equipped with ``black boxes'' 
(i.e., EDRs) analogous to those found on commercial aircraft.
---------------------------------------------------------------------------

    \11\ Docket No. NHTSA-1998-4367-22.
---------------------------------------------------------------------------

    In responding to these petitions, NHTSA acknowledged that EDRs 
could provide valuable information useful for analyzing crashes and 
improving motor vehicle safety. However, the agency decided to deny the 
petitions because the motor vehicle industry was already voluntarily 
moving in the direction recommended by the petitioners, and because the 
agency believed ``this area presents some issues that are, at least for 
the present time, best addressed in a non-regulatory context.'' \12\
---------------------------------------------------------------------------

    \12\ 63 FR 60270, 60270 (Nov. 9, 1998) (Docket No. NHTSA-1998-
4672-1); 64 FR 29616, 29616 (June 2, 1999) (Docket No. NHTSA-1999-
5737-1).
---------------------------------------------------------------------------

2. Petition From Dr. Ricardo Martinez
    In October 2001, the agency received a petition \13\ from Dr. 
Ricardo Martinez, President of Safety Intelligence Systems Corporation 
and former Administrator of NHTSA, asking us to ``mandate the 
collection and storage of onboard vehicle crash event data, in a 
standardized data and content format and in a way that is retrievable 
from the vehicle after the crash.''
---------------------------------------------------------------------------

    \13\ Docket No. NHTSA-2002-13546-3.
---------------------------------------------------------------------------

    In his petition for rulemaking, Dr. Martinez argued that 
understanding what happens in a crash is essential to preventing 
injuries and deaths, and that EDRs would improve crash reconstruction 
analysis. The petitioner also stated that current crash reconstruction 
analysis is costly, time consuming, laborious, and often inaccurate. 
According to Dr. Martinez, the increasing sophistication and decreasing 
costs of information technology have created the opportunity to now 
mandate the capture, storage, and retrieval of onboard crash data, and 
a NHTSA rulemaking could greatly accelerate the development of ACN.
    The petition from Dr. Martinez was submitted shortly after the 
NHTSA EDR Working Group had published its final report. As discussed in 
more detail in the next section of this document, in October 2002, 
after the second working group had completed its work, we decided to 
request public comments on what future role the agency should take 
related to the continued development and installation of EDRs in motor 
vehicles. We decided to respond to Dr. Martinez's petition after 
considering those comments.

D. October 2002 Request for Comments

    On October 11, 2002, NHTSA published a request for comments 
concerning EDRs in the Federal Register (67 FR 63493).\14\ In that 
document, the agency discussed its prior involvement concerning EDRs, 
and it requested comments on what future role NHTSA should take related 
to the continued development and installation of EDRs in motor 
vehicles. The request for comments discussed a range of issues, 
including safety benefits, technical issues, privacy issues, and the 
role of the agency, and it also posed several questions.
---------------------------------------------------------------------------

    \14\ Docket No. NHTSA-2002-13546-1.
---------------------------------------------------------------------------

    In response to this request, we received comments from light and 
heavy vehicle manufacturers, equipment manufacturers, vehicle users, 
the medical community, advocacy organizations, safety research 
organizations, crash investigators, insurance companies, academics, and 
government agencies. We also received comments from a number of private 
individuals.
    To summarize, these comments raised issues concerning the safety 
benefits of EDRs (with most commenters suggesting EDRs will improve 
vehicle safety), technical issues surrounding a potential rulemaking on 
EDRs (such as the types of data elements to be collected, amount of 
data to be recorded, and crash survivability of EDR data), potential 
privacy issues associated with EDRs, NHTSA's role in the future of 
EDRs, and public perception of EDRs.
    After considering the comments and other information NHTSA had 
gathered on EDRs, NHTSA decided to grant the Martinez petition in part 
and commenced rulemaking.

III. Notice of Proposed Rulemaking

A. Summary of the NPRM

    On June 14, 2004, NHTSA published a NPRM in the Federal Register 
(69 FR 32932)\15\ proposing to: (1) Require that EDRs voluntarily 
installed in light vehicles record a minimum set of specified data 
elements useful for crash investigations, analysis of safety equipment 
performance, and automatic

[[Page 51005]]

collision notification systems; (2) specify requirements for data 
format; (3) increase the survivability of the EDRs and their data by 
requiring that the EDRs function during and after the front, side, and 
rear vehicle crash tests specified in several Federal motor vehicle 
safety standards; (4) require vehicle manufacturers to make publicly 
available information for a download tool that would enable crash 
investigators to retrieve data from the EDR; and (5) require vehicle 
manufacturers to include a brief standardized statement in the owner's 
manual indicating that the vehicle is equipped with an EDR and 
describing purposes of that device.
---------------------------------------------------------------------------

    \15\ Docket No. NHTSA-2004-18029-2.
---------------------------------------------------------------------------

    NHTSA tentatively concluded that the proposed requirements would 
help ensure that EDRs record, in a readily usable manner, the data 
necessary for effective crash investigations, analysis of safety 
equipment performance, and automatic crash notification systems. NHTSA 
stated its belief that its proposal would help provide a better 
understanding of the circumstances under which crashes and injuries 
occur and would lead to derivative benefits, such as safer vehicle 
designs.
    In the NPRM, NHTSA responded to the Martinez petition \16\ for 
rulemaking, which asked the agency to ``mandate the collection and 
storage of onboard vehicle crash event data, in a standardized data and 
content format and in a way that is retrievable from the vehicle after 
the crash.'' The agency granted the petition in part, to the extent 
that it proposed a regulation to specify standardized data content and 
format for EDRs in a manner that is retrievable from a vehicle after a 
crash.
---------------------------------------------------------------------------

    \16\ Docket No. NHTSA-2002-13546-3.
---------------------------------------------------------------------------

    However, NHTSA denied the petition to the extent that the agency 
did not propose to mandate EDRs. In the NPRM, the agency stated its 
belief that a mandatory EDR rule was not the best approach at this 
time, and we noted that the industry is continuing to move in the 
direction of installing EDRs in an increasing percentage of new 
vehicles. Further, the industry trend is toward designing EDRs to 
include greater amounts of crash data. Given this trend, we did not 
deem it necessary for us to propose to require the installation of 
EDRs, but remained open to considering this in the future.
    The NPRM also discussed other key issues including data elements to 
be recorded, data standardization, data retrieval, crash survivability, 
privacy, and lead time. The NPRM provided detailed tables of the data 
elements to be recorded under the proposal and the relationship of the 
data elements to the stated purposes of the rulemaking. While the NPRM 
did propose specific technical requirements and specifications, NHTSA 
requested comments on the proposed data elements, including whether the 
list sufficiently covers technology that is likely to be in vehicles in 
the next five to ten years.
    In terms of data standardization, the NPRM proposed a standardized 
format for each data element, specifying the corresponding recording 
intervals/times, units of measurement, sampling rates, data range/
accuracy/precision requirements, and where appropriate, filter class. 
However, the NPRM noted that there was currently not an industry 
standard for EDR format.
    The NPRM also solicited comments on EDR data retrieval. 
Specifically, NHTSA sought alternative approaches to the data retrieval 
requirements proposed in the NPRM, which would have required vehicle 
manufacturers to submit specifications for accessing and retrieving the 
stored EDR data and information in sufficient detail to permit 
companies that manufacture diagnostic tools to develop and build 
devices for accessing and retrieving the EDR's stored data.
    Regarding the functioning of EDRs and crash survivability, the NPRM 
proposed requirements for the EDR trigger threshold, EDR recording in 
multi-event crashes, capture of EDR data, and the performance of EDRs 
in crash tests.
    The NPRM discussed privacy issues related to EDRs, but it also 
noted that most privacy issues involve Federal and State laws separate 
from NHTSA's primary statutory authority.
    Finally, the NPRM discussed lead time for the regulation's proposed 
compliance date. The NPRM proposed a compliance date of September 1, 
2008, to permit manufacturers to make EDR-related design changes as a 
part of their regular production cycle in order to minimize costs.

B. Summary of Public Comments to the NPRM

    NHTSA received over 100 comments on the NPRM from automobile 
manufacturers,\17\ motor vehicle equipment suppliers and 
businesses,\18\ trade associations,\19\ advocacy and special interest 
groups,\20\ and individuals. (All of the comments on the NPRM can be 
reviewed in Docket No. NHTSA-2004-18029.) Commenters expressed a wide 
range of views, with vehicle manufacturers, motor vehicle equipment 
suppliers, and trade associations generally supporting the NPRM in 
concept, while raising a number of significant issues and recommending 
modifications. Special interest groups advocating highway safety 
generally called for a more extensive regulation; for example, these 
commenters asked NHTSA to require EDRs in all vehicles, to require more 
data elements to be recorded, and/or to require uniform EDR data 
retrieval so that first responders and other emergency personnel may 
easily access EDR data. A number of individuals who commented on the 
NPRM raised potential privacy concerns.
---------------------------------------------------------------------------

    \17\ Comments were received from the following vehicle 
manufacturers: (1) American Honda Motor Company (Honda); (2) 
DaimlerChrysler, VSO (DaimlerChrysler); (3) Ford Motor Company 
(Ford); (4) General Motors Corporation (GM); (5) Hyundai America 
Technical Center, Inc. (Hyundai and Kia); (6) Mitsubishi Motors R & 
D of America, Inc. (Mitsubishi); (7) Nissan North American, Inc. 
(Nissan); (8) Porsche Cars North American, Inc. (Porsche); (9) 
Subaru of America, Inc. (Subaru); and (10) Toyota Motor North 
America, Inc. (Toyota).
    \18\ Comments were received from the following motor vehicle 
equipment suppliers and other businesses: (1) Bendix Commercial 
Vehicle Systems, L.L.C. (Bendix); (2) Delphi; (3) Gelco Corporation 
d/b/a GE Fleet Services (Gelco); (4) Kast, GmbH (Kast); (5) Injury 
Sciences, L.L.C. (Injury Sciences); (6) Racing Information Systems; 
(7) Safety Intelligence Systems Corporation (SISC); (8) Siemens VDO 
Automotive, AG (Siemens); (9) TRW Automotive (TRW); and (10) Wyle 
Laboratories, Inc. (Wyle Laboratories).
    \19\ Comments were received from the following trade 
associations: (1) Alliance of Automobile Manufacturers (Alliance); 
(2) American Trucking Association (ATA); (3) Association of 
International Automobile Manufacturers, Inc.--Technical Affairs 
Committee (AIAM); (4) National Automobile Dealers Association 
(NADA); (5) Property Casualty Insurers Association of America 
(PCIAA); and (6) Specialty Equipment Market Association (SEMA).
    \20\ Comments were received from the following advocacy (and 
other) groups: (1) Advocates for Highway and Auto Safety 
(Advocates); (2) Albemarle County Police Department; (3) American 
Automobile Association (AAA); (4) Canada Safety Council; (5) 
Children's Hospital of Philadelphia; (6) Electronic Privacy 
Information Center (EPIC); (7) European Commission; (8) Garthe 
Associates (Garthe); (9) Institute of Electrical and Electronics 
Engineers Vehicular Technology Society (IEEE-VTS); (10) Insurance 
Institute for Highway Safety (IIHS); (11) National Motorist 
Association; (12) National Transportation Safety Board (NTSB); (13) 
Public Citizen; and (14) Society of Automotive Engineers (SAE).
---------------------------------------------------------------------------

    The following overview of the public comments reflects the key 
issues raised by the commenters, including whether the EDR rule should 
be mandatory, the number and type of data elements to be recorded, EDR 
data standardization requirements, EDR data retrieval and whether to 
require a standardized data retrieval tool/universal interface, and EDR 
crash survivability. Other commenters addressed the proposed owner's 
manual disclosure statement, potential privacy concerns, lead time, and 
costs. A more in-depth analysis of

[[Page 51006]]

comments along with the agency's response follows in section IV.B of 
this document.
Whether NHTSA Should Require EDRs
    In their comments, most automobile manufacturers supported the EDR 
standardization requirements for voluntarily-installed EDRs. However, 
GM, Ford, some industry associations, and most advocacy and special 
interest groups, urged NHTSA to require EDRs to be installed in all new 
vehicles. Commenters as diverse as GM and Public Citizen urged 
mandatory installation of EDRs. Arguments for why installation should 
be mandatory varied, but included concerns that manufacturers will opt 
out under a voluntary installation approach, that standardization 
requirements for voluntary-installed EDRs will discourage EDR 
installation, and that voluntary installation would take many years to 
build up sufficient information for useful study.
Number and Types of Required Data Elements
    The NPRM separated EDR data elements into two categories. The first 
category consisted of a set of data elements that must be recorded if 
an automobile manufacturer currently uses an EDR for any one data 
element (i.e., ``required'' data elements). The second category 
consisted of data elements that must be recorded only if the vehicle is 
equipped with a specified system or sensing capability (i.e., ``if 
equipped'' data elements). The NPRM listed 18 required data elements 
and an additional 24 ``if equipped'' data elements.
    Overall, automobile manufacturers, and other commenters connected 
to the automotive industry, stated their belief that the number of 
proposed required data elements is excessive in light of NHTSA's stated 
purposes. However, manufacturers differed in their assessment as to 
which of the data elements should be required to be recorded and their 
rationale why. The manufacturers agreed that the number of data 
elements should be reduced due to: (1) The estimated (excessive) cost 
of the EDR proposal; (2) limitations in memory and microprocessing 
capability of EDRs; (3) the potential to inhibit collection of more 
useful data; and (4) the desire to avoid complete electrical redesigns.
    In contrast, highway safety advocacy groups, such as Public Citizen 
and Advocates, suggested that the number of required elements is 
insufficient. This group of commenters generally argued that more data 
elements should be recorded in order to: (1) Provide additional data 
contribution for a more definitive crash causation evaluation; (2) 
address equipment likely to be used in the future; and/or (3) encourage 
uniformity. Some commenters, including Injury Sciences and Public 
Citizen, suggested adding the Vehicle Identification Number (VIN) as a 
recorded data element. Still others commented that certain data 
elements in the ``if equipped'' category should be moved to the 
``required'' category or vice versa.
EDR Data Standardization
    The NPRM proposed specific technical specifications for each data 
element, including sampling rates and recording intervals, data 
standardization requirements, and data retrieval requirements. The 
commenters on this issue, mostly from the automobile industry, raised 
concerns about the proposed recording frequency and sampling rates, 
especially regarding the amount of microprocessing and memory required 
to process and store the proposed EDR data. According to the 
manufacturers, the increase in microprocessing and memory capabilities 
that would be required to comply with the proposed rule would be more 
costly than the agency anticipated. Therefore, manufacturers 
recommended alternative sampling rates and recording intervals that 
they believe would be less expensive. Automotive industry commenters 
also recommended other technical adjustments to the proposed recording 
requirements. They also generally disagreed with the proposed multiple-
event recording requirement, with most stating that it is unnecessary 
and not current industry practice.
    Automobile manufacturers generally commented that the range, 
accuracy, and precision specifications contained in the NPRM should not 
be included in the final rule because the proposed parameters are 
beyond what is currently utilized in the state-of-the-art EDRs and the 
provisions are not necessary to achieve the agency's goals. Other 
commenters agreed with the concept of standardization, but suggested 
that it be accomplished in another manner, such as leaving it to the 
discretion of the manufacturers for optimal restraint system 
performance or applying SAE J1698.
    Highway safety advocates commented that sampling rates and 
recording intervals should be of sufficient duration to record the full 
crash event, especially for ``rollover'' crashes.
EDR Data Retrieval and Whether To Require a Standardized Data Retrieval 
Tool
    With regard to data retrieval requirements, most manufacturers 
objected to furnishing non-proprietary technical specifications to 
NHTSA and offered alternative approaches for retrieving EDR data, such 
as through licensing agreements or making retrieval tools available to 
the public at a reasonable price. Highway safety advocacy groups argued 
that NHTSA should require standardization of data retrieval methods, 
that first responders should have access to EDR data, and that NHTSA 
should require a uniform architecture for data retrieval with a 
standardized interface location.
EDR Survivability and Crash Test Performance Requirements
    The NPRM proposed that EDR data must exist upon completion of each 
crash test and be retrievable by a methodology specified by the vehicle 
manufacturer for not less than 30 days after the test and without 
external power. Several automobile manufacturers commented that the 
proposed crash test requirement is impracticable because they believe 
it would require tests to be performed with engines running and various 
vehicle systems activated, which would cause a danger to test 
personnel. As an alternative, commenters suggested a simulated 
laboratory test. Automobile manufacturers commented that the proposed 
rule would greatly increase testing costs. There were also comments on 
whether an alternative power source would be required to meet the 30-
day provision in Sec. 563.10(d). Other commenters, including NTSB and 
Public Citizen commented that NHTSA should require that EDR data 
survive fire, fluid immersion, and severe crashes.
Other Issues
    The NPRM proposed a compliance date of September 1, 2008, for the 
EDR regulation. Nearly all commenters, especially automobile 
manufacturers, believed that the agency underestimated the amount of 
time needed to meet the requirements of the proposed rule. Several 
manufacturers suggested that, as part of the final rule, the agency 
should provide a phase-in (e.g., a four-year phase-in beginning in 
2008).
    In order to educate the public about EDRs and to gain public 
acceptance for use in passenger vehicles, the NPRM proposed that 
vehicles equipped with an EDR must also include a specified statement 
in the owner's manual. This

[[Page 51007]]

statement would inform the vehicle owner about the presence of the EDR 
and its purposes. Most commenters, including automobile manufacturers 
and privacy advocates, expressed support for a disclosure statement. 
However, several commenters (including automobile manufacturers, EPIC, 
and individuals) suggested alternative language. Comments concerning 
the disclosure statement ranged from concerns about privacy and 
ownership of the EDR data to preemption and State disclosure 
requirements.
    Commenters, especially individuals, raised concerns about privacy. 
In the NPRM, we addressed privacy issues, stating our position that 
NHTSA's use of the data collected from EDRs would not raise privacy 
concerns. NHTSA obtains the owner's consent for collecting and using 
EDR data and carefully protects any information that could potentially 
be used to identify an individual. In the context of EDRs, the 
information in question that may be linked to an individual is the 
vehicle identification number (VIN), which is collected at the time EDR 
information is downloaded. The following discussion explains why it is 
necessary for the agency to collect VIN information in connection with 
EDRs, how such information is used by the agency, and the safeguards 
the agency takes related to the release of such information.
    VIN information (e.g., relevant to the make/model in question) is 
necessary to download and process the EDR data, because the commercial 
EDR download tool requires the VIN to be inputted into the program in 
order to link the EDR file with data to ensure proper engineering 
output. Without VIN input, similar data may mean different things 
depending on the vehicle from which it comes.
    This final rule does not require EDRs to record VIN information. 
However, the full VIN of a vehicle must be inputted into current EDR 
extraction tools as a key to ensure proper conversion of the electronic 
EDR data to a usable format. The full VIN is needed in order to account 
for running changes that may occur during a particular model year, 
thereby rendering it infeasible to use a shortened VIN. We note that 
such VIN information is normally available through other means during 
the course of crash reconstruction (i.e., through reading the VIN label 
on the vehicle itself). Further, other parties, such as law 
enforcement, could combine the EDR data with the type of personally 
identifying data routinely acquired during a crash investigation.
    In terms of the use of EDR data, the agency takes the EDR-generated 
information that it collects and incorporates the information into 
large crash-related databases in order to gain a more comprehensive 
understanding of certain crash events; the information contained in 
these databases is not retrieved or retrievable by name or other 
individual identifier.
    The agency's rationale for protection of the VIN information 
contained in EDRs is as follows. By way of background, the VIN data 
identify the vehicle itself and do not specifically provide name, 
address, or other personal identifier information on an individual. 
Furthermore, EDR data alone cannot confirm exactly who was driving the 
vehicle at any given time (e.g., vehicle owner or other individuals 
(either with or without permission)). However, even though VIN 
information is not a ``record'' \21\ or part of a ``system of records'' 
\22\ as those terms are defined under the Privacy Act, NHTSA has 
nevertheless taken steps to prevent the release of VIN information, 
because VIN information can be used in various commercially-available 
programs to determine the identity of the current owner of a vehicle.
---------------------------------------------------------------------------

    \21\ The Privacy Act of 1974 defines ``record'' as ``any item, 
collection, or grouping of information about an individual that is 
maintained by an agency, including but not limited to, his 
education, financial transactions, medical history, and criminal or 
employment history and that contains his name, or the identifying 
number, symbol, or other identifying particular assigned to the 
individual, such as a finger or voice print or a photograph.'' 5 
U.S.C. 552(a)(4).
    \22\ The Privacy Act defines ``system of records'' as ``a group 
of any records under the control of any agency from which 
information is retrieved by the name of the individual or by some 
identifying number, symbol, or other identifying particular assigned 
to the individual.'' 5 U.S.C. 552(a)(5).
---------------------------------------------------------------------------

    As a practical matter, information contained in these records that 
has the potential indirectly to identify individuals is not made 
public, except as specifically required by law. Furthermore, prior to 
the release of information from databases containing EDR data (usually 
aggregated reports), the agency strips out the last six characters of 
the VIN (i.e., the portion that would allow identification of a 
specific vehicle and, potentially by indirect means, the identity of 
the vehicle's current owner). In light of the above, we believe that 
the agency has taken adequate steps to ensure individual privacy vis-
[agrave]-vis its use of EDR data.
    However, we recognized that there may be privacy issues associated 
with EDRs related to the use of EDR data by entities other than NHTSA, 
such as law enforcement and EMS personnel, other government entities, 
and the automotive industry. Notwithstanding our extensive treatment of 
the privacy issue, we still received comments from individuals who 
believe that EDRs are an intrusion of their privacy because EDRs might 
record aspects of their driving behavior (e.g., whether they are 
speeding or not wearing a safety belt) that they do not want to be 
known. Automobile manufacturers and highway safety groups commented 
that the potential benefits of EDRs outweigh any privacy concerns.
    In addition to lead time, privacy, and owner's manual disclosure 
statement issues, commenters raised additional substantive issues 
including cost, preemption, and inclusion of ACN as a goal of the EDR 
rule.
    Many commenters, mostly automobile manufacturers, believed that 
NHTSA's cost estimates were significantly understated. According to 
these commenters, the proposed requirements outlined in the NPRM would 
contribute to higher costs because of the additional microprocessors 
and memory needed to handle larger amounts of saved data. These 
commenters also argued that the dynamic testing requirements would 
increase costs along with the requirements of accuracy, range, and 
precision, which they argued are in excess of current industry 
practice.
    Commenters requested that NHTSA specifically preempt inconsistent 
State and local regulations related to EDRs. Automobile manufacturers 
were concerned about the possibility of having to comply with 
inconsistent State regulations, especially concerning owner's manual 
disclosure statements and technical specifications of EDRs.
    With respect to ACN as a stated goal of the EDR rule, commenters 
associated with the automotive industry argued that this goal should be 
removed, since the proposed rule would not require ACN or specifically 
state that the rule will not limit the ability of manufacturers to 
offer ACN.
    Other, more specific and technical issues were raised by 
commenters. These issues will be treated and addressed in section IV.B 
of this notice.

IV. The Final Rule and Response to Public Comments

A. The Final Rule

1. Summary of the Requirements
    After careful consideration of the public comments on the NPRM, we 
are promulgating this final rule to establish a regulation for 
voluntarily-installed EDRs in order to standardize EDR data. The 
requirements of this regulation are tailored to advance the stated 
purposes of this rulemaking without imposing unnecessary burdens or 
unduly

[[Page 51008]]

impeding the future technological development of EDRs. In overview, the 
final rule specifies uniform, national requirements for EDR-equipped 
vehicles covered by the regulation, including the collection, storage, 
and retrievability of onboard motor vehicle crash event data. It also 
specifies requirements for vehicle manufacturers to make retrieval 
tools and/or methods commercially available so that crash investigators 
and researchers are able to retrieve data from EDRs.
    Specifically, the regulation applies to passenger cars, 
multipurpose passenger vehicles, trucks, and buses with a GVWR of 3,855 
kg (8,500 pounds) or less and an unloaded vehicle weight of 2,495 kg 
(5,500 pounds) or less, except for walk-in van-type trucks or vehicles 
designed to be sold exclusively to the U.S. Postal Service, that are 
equipped with an event data recorder and to manufacturers of these 
vehicles.\23\ Subject to an exception for final-stage manufacturers and 
alterers discussed below, compliance with the requirements of the final 
rule commences for covered vehicles manufactured on or after September 
1, 2010. The final rule is intended to be technology-neutral, so as to 
permit compliance with any available EDR technology that meets the 
specified performance requirements.
---------------------------------------------------------------------------

    \23\ These are the same applicability limits set for the air bag 
requirements in frontal crashes in Federal Motor Vehicle Safety 
Standard No. 208.
---------------------------------------------------------------------------

    The following points highlight the key provisions of the final 
rule:
     Term ``event data recorder'' is defined as ``a device or 
function in a vehicle that captures the vehicle s dynamic, time-series 
data during the time period just prior to a crash event (e.g., vehicle 
speed vs. time) or during a crash event (e.g., delta-V vs. time), such 
that the data can be retrieved after the crash event. For the purposes 
of this definition, the event data do not include audio and video 
data.''
     Each vehicle equipped with an EDR must record all of the 
data elements listed in Table I, during the interval/time and at the 
sample rate specified in that table. There are 15 required data 
elements (see paragraph 563.7(a), Table I). Examples of these data 
elements are ``delta-V, longitudinal,'' ``maximum delta-V, 
longitudinal,'' ``speed, vehicle indicated,'' and ``safety belt status, 
driver.''
     Each vehicle equipped with an EDR that records any of the 
data elements listed in Table II identified as ``if recorded'' (most 
elements in that table) must capture and record that information 
according to the interval/time and at the sample rate specified in that 
table. Data elements listed in Table II as ``if equipped'' (i.e., 
``frontal air bag deployment, time to nth stage, driver'' and ``frontal 
air bag deployment, time to nth stage, right front passenger'') must 
record the specified information, even if they are not presently doing 
so. (The ``frontal air bag deployment, time to nth stage'' data 
elements provide critical timing data for vehicles equipped with multi-
stage air bags, which will help in assessing whether an air bag is 
deploying correctly during a crash (i.e., whether the sensors are 
functioning properly). In drafting this final rule, we had considered 
including these two elements as required elements under Table I, but we 
recognized that not all vehicles are equipped with multi-stage air 
bags. Thus, by including these elements in Table II and requiring 
recording of that information if the vehicle is so equipped, we are, in 
effect, requiring this data from all vehicles equipped with an EDR and 
multi-stage air bags.)
    There are 30 data elements included in Table II (see paragraph 
563.7(b), Table II). Examples of these data elements are ``lateral 
acceleration,'' ``longitudinal acceleration,'' ``frontal air bag 
suppression switch status, right front passenger (on, off, or auto),'' 
``frontal air bag deployment, time to nth stage, driver,'' and ``safety 
belt status, right front passenger (buckled, not buckled).''
     The data elements required to be collected by the EDR 
pursuant to Tables I and II, as applicable, must be recorded in 
accordance with the range, accuracy, and resolution requirements 
specified in Table III, Recorded Data Element Format (see paragraph 
563.8(a), Table III).
     For EDRs that record acceleration, the longitudinal, 
lateral, and normal acceleration time-history data must be filtered in 
accordance with the filter class specified in Table III (i.e., SAE 
J211-1, Class 60) (see paragraph 563.8(b)). Such filtering may be done 
during collection or post-processing.
     The EDR must collect and store data elements for events in 
accordance with the following conditions and circumstances as specified 
in paragraph 563.9:
    (1) In an air bag deployment crash, the data recorded from any 
previous crash must be deleted; the data related to the deployment must 
be recorded, and the memory must be locked in order to prevent any 
future overwriting of these data.
    (2) In an air bag non-deployment crash that meets the trigger 
threshold, all previously recorded data in the EDR's memory must be 
deleted from the EDR's memory, and the current data (up to two events) 
must be recorded.
     In order to ensure that survivability of EDR data in most 
crashes, the EDR is tested in conjunction with crash tests already 
required under FMVSS No. 208, Occupant Crash Protection, and FMVSS No. 
214, Side Impact Protection (see paragraph 563.10). Except for the 
elements discussed below, the data elements required under paragraph 
563.7 must be recorded in the format specified by paragraph 563.8, must 
exist at the completion of the crash test, and must be retrievable by 
the methodology specified by the vehicle manufacturer (as required 
under paragraph 563.12) for not less than 10 days after the test. The 
``complete file recorded (yes, no)'' data element must read ``yes'' 
after the test.
    The EDR need not meet the above survivability requirements for the 
following data elements: (1) ``Engine throttle, % full,'' (2) ``service 
brake, on/off,'' and (3) ``engine RPM.'' These elements have been 
excluded from these requirements because vehicles are crash tested 
without the engine running for safety reasons, so the EDR would not be 
able to record the above data elements under those circumstances.
     For vehicles equipped with an EDR, vehicle manufacturers 
must include a specified statement in the owner's manual to make the 
operator aware of the presence, function, and capabilities of the EDR 
(see paragraph 563.11).
     In order to ensure the retrievability of EDR data, each 
vehicle manufacturer that installs EDRs must ensure by licensing 
agreement or other means that retrieval tool(s) are commercially 
available for downloading the required EDR data. The retrieval tool 
must be commercially available not later than 90 days after the first 
sale of the vehicle for purposes other than resale.
2. Lead Time
    In order to minimize the costs associated with the standardization 
of EDR data, we have stated our intention to provide adequate lead time 
to manufacturers to enable them to incorporate necessary changes as 
part of their routine production cycles. In the NPRM, we had proposed a 
compliance date of September 1, 2008. However, in their comments on our 
proposal, some manufacturers had argued that a longer lead time is 
needed to make the necessary design and production changes. Others 
requested a phase-in of the EDR requirements, which was characterized 
as particularly important for manufacturers that already have a

[[Page 51009]]

significant portion of their fleet equipped with EDRs.
    In light of the fact that installation of EDRs remains voluntary on 
the part of vehicle manufacturers and our concomitant desire to 
minimize costs, we have decided to adopt the recommendations of 
commenters to provide vehicle manufacturers with additional lead time. 
Accordingly, subject to the exception below, we have decided to require 
covered vehicles manufactured on or after September 1, 2010 to comply 
with the requirements of this final rule. We believe that lead time in 
excess of four years, particularly given the revised technical 
requirements, should prove adequate for all vehicle manufacturers and 
all vehicle lines, without the need for a phase-in. Vehicle 
manufacturers may voluntarily comply with these requirements prior to 
this date.
    Beyond the suggestions of the automobile manufacturers to increase 
the lead time associated with this rule, NHTSA conducted its own 
analysis of the technical changes needed to meet the standardization 
requirements and specifications of this final rule. As discussed below, 
we determined that the final rule will necessitate a number of design 
and technical changes to current EDRs.
    For example, current EDR systems have been designed independently 
by the vehicle manufacturers, thereby resulting in differences in data 
definitions. Thus, in implementing this final rule, manufacturers will 
need to make technical changes to their systems to reflect 
standardization in the data elements.
    Further, we have added new definitions related to EDR operation 
that will necessitate changes to EDRs. The ``trigger threshold'' 
required by this final rule is different than that which any vehicle 
manufacturer currently utilizes. Generally, vehicle manufacturers use 
wake-up levels to start collecting data, based upon vehicle 
deceleration. However, our final rule specifies that data collection be 
triggered by using change-in-velocity (delta-V) over a specified time 
period, which will require algorithm development and possibly 
additional non-volatile memory buffers to capture and analyze these 
vehicle data. The two-event capture and recording requirement in the 
final rule is also different from that which any vehicle manufacturer 
currently uses. While some current EDRs do capture and record two 
events, the data are not captured with standardized logic, as is 
specified in the final rule (e.g., standardization of the calculation 
of time between events). Another new requirement is that the EDR must 
lock the file if an air bag deploys during an event; this requirement 
is one that will need to be newly implemented by most of vehicle 
manufacturers.
    Another requirement in the final rule that is likely to necessitate 
changes in EDRs is the requirement for the capture and recording of 
pre-crash data. With the exception of GM and Toyota, no other vehicle 
manufacturer captures and records pre-crash data that can be downloaded 
using a commercially available tool. Ford is developing a pre-crash 
data recording capability, but Ford is collecting those data in the 
engine control module. All other vehicle manufacturers will need to 
update their systems to achieve pre-crash data collection, which will 
necessitate algorithm development and possibly additional non-volatile 
memory to continuously capture and hold these data until an event 
occurs. Further, the sampling of the pre-crash data will need to be 
standardized to two samples per second, in order to meet the 
requirements of the final rule. To our knowledge, no vehicle 
manufacturer currently collects pre-crash data at this sample rate 
(e.g., most GM and Toyota vehicles capture data at one sample per 
second during the interval specified in the final rule). Again, 
updating these systems in this fashion will require additional 
algorithm development and possible additional non-volatile memory.
    In addition, we anticipate that development of a turnkey operation 
for downloading EDR data will take significant time to accomplish. 
Vehicle manufacturers will need time to develop their licensed partner 
relationships for production of download tools.
    Finally, we note that the latest version of GM's EDR (e.g., ones 
used in the 2004 Malibu) does not capture and record delta-V data 
within the accuracy requirement specified in the final rule. In two 
tests performed by IIHS, which shared results with NHTSA for use in a 
paper for presentation at an International Technical Conference on the 
Enhanced Safety of Vehicles (ESV), the delta-Vs recorded by the EDR 
were at or outside the accuracy specifications of the final rule.\24\ 
Additionally, we note that GM has previously reported that the current 
generation of EDRs have data resolution and accuracy outside the levels 
specified in the final rule.\25\ In sum, sufficient lead time will be 
required for vehicle manufacturers to make the changes necessitated by 
the final rule without incurring significant additional costs.
---------------------------------------------------------------------------

    \24\ Niehoff, Peter, et al., Evaluation of Event Data Recorders 
in Full Systems Crash Tests, ESV Paper No. 05-0271 (2005).
    \25\ ``Recording Automotive Crash Event Data,'' Chidester, 
Hinch, Mercer & Schultz, NTSB (1999). See http://www.ntsb.gov/events/symp_rec/proceedings/authors/chidester.pdf.
---------------------------------------------------------------------------

    Consistent with the policy set forth in NHTSA's February 14, 2005 
final rule on certification requirements under Federal motor vehicle 
safety standards for vehicles built in two or more stages and altered 
vehicles (70 FR 7414), final-stage manufacturers and alterers of 
covered vehicles must comply with the requirements of this final rule 
for vehicles manufactured on or after September 1, 2011. However, 
final-stage manufacturers and alterers may voluntarily comply with the 
requirements of the regulation prior to this date.

B. Response to Public Comments

1. Whether NHTSA Should Require EDRs
    We expressly stated in the NPRM that we were not proposing to 
require all light vehicles to be equipped with EDRs. Under the proposed 
rule, vehicle manufacturers retained discretion regarding the decision 
of whether to install EDRs. However, if a vehicle were equipped with an 
EDR, the vehicle would be required to comply with the requirements of 
the proposed Part 563. We stated that we did not believe it was 
necessary to mandate installation of EDRs at this time, noting that the 
industry has substantially progressed in the development and 
installation of EDRs without the agency's requiring them. We estimated 
that at least 64 percent of model year 2004 passenger cars and other 
light vehicles have some recording capability, and more than half 
record elements such as crash pulse data. We noted also that industry 
was expected to install EDRs in an increasing percentage of new 
vehicles.
    The agency received several comments on the issue of whether we 
should require manufacturers to install EDRs in all new vehicles. GM 
commented that NHTSA should adopt a FMVSS that would mandate 
installation of EDRs on all passenger cars and light trucks with a GVWR 
up to 3,855 kg (8,500 pounds). GM stated that a mandatory EDR 
requirement would maximize safety benefits by ensuring that all covered 
vehicles capture and record key crash data. According to GM, an EDR 
mandate would also eliminate incentives for manufacturers to remove 
existing EDRs or to delay their introduction. In addition, GM argued 
that the standard should prohibit

[[Page 51010]]

switches that would permit EDR disablement.
    Public Citizen, Advocates, NADA, and NTSB urged NHTSA to require 
the installation of EDRs. Public Citizen stated that NHTSA should 
require EDRs because these devices can provide valuable safety 
benefits, including: (1) Better understanding of crash causation and 
injury sources; (2) enhanced commercial vehicle safety; (3) better data 
on defect trends; (4) safer highway designs; and (5) improved emergency 
response to crashes. Advocates argued that unless the agency requires 
EDRs, data collection would take many years to gather sufficient 
information for useful study. Mr. Fink, a crash reconstructionist, 
stated that the rule should require EDRs in all vehicles sold in the 
U.S. Four individuals commented that they supported the NPRM, one of 
which indicated that EDRs should be mandatory.
    Several commenters argued that NHTSA's proposal to apply the rule 
to only those vehicles equipped with EDRs would either act as an 
incentive for manufacturers to remove EDRs from product lines currently 
equipped with EDRs or would discourage manufacturers from installing 
EDRs in new product lines. Ford argued the agency would need to issue a 
rule that requires installation of EDRs to accomplish the objectives 
set forth in the agency's proposal. Ford stated that it has been unable 
to develop a workable definition of an EDR that would uniformly create 
a truly voluntary requirement for all vehicle manufacturers and that 
avoids incentives for removal of existing recording capability or the 
deferred introduction of such capabilities.
    IIHS, Public Citizen, PCIAA, and Children's Hospital of 
Philadelphia joined Ford and GM in arguing that not requiring 
manufacturers to install EDRs would act as an incentive for vehicle 
manufacturers to remove EDRs from vehicles and/or would discourage 
installation of EDRs in new product lines. According to these 
commenters, the net result would be a reduction in the number of 
vehicles equipped with EDRs. While Ford expressed support for modifying 
language to create a truly voluntary requirement that would at the same 
time address these concerns, IIHS, Public Citizen, PCIAA, GM, and 
Children's Hospital argued for a mandatory rule (with PCIAA noting that 
the industry needs ample lead time to comply), which would eliminate 
the incentive to remove EDRs and/or the discretion not to install EDRs 
in new product lines.
    SISC supported the proposal's position that EDRs should include 
minimum standards for capturing crash data. SISC stated that current 
EDRs are focused on capturing data to evaluate the performance of 
safety systems; however, they do not adequately address the needs of 
capturing data for crash investigations. SISC stated that without 
mandatory minimum standards for capturing crash data, EDRs would not 
provide the type of information needed for safety research.
    On the other hand, DaimlerChrysler and Toyota supported a voluntary 
approach to EDR installation. DaimlerChrysler also commented that the 
definition of EDR should be modified to ensure that EDRs are voluntary. 
In explaining its request for modification, DaimlerChrysler stated that 
the NPRM's definition of EDR references the deployable restraint 
control module for the purpose of determining whether a vehicle is 
equipped with an EDR. DaimlerChrysler argued that all light vehicles 
are equipped with such control modules; therefore, the adoption of a 
definition making such a reference would effectively mandate EDRs for 
all applicable vehicles, contrary to the agency's stated intent.
    Porsche also argued that the NPRM's definition of EDR would 
effectively require manufacturers to install EDRs. Porsche argued that 
a vehicle might be capable of recording and storing a few pieces of 
static freeze frame data in the air bag control unit (i.e., an isolated 
observation or snapshot of a set of data such as the seat belt status, 
frontal air bag warning lamp status, etc., triggered by an impact 
exceeding a defined trigger threshold). Although such systems fall 
outside the common understanding of EDRs, Porsche argued that this type 
of recorded data would fall within the proposed EDR definition. Porsche 
stated that storage of freeze frame data should not, by itself, be a 
sufficient basis for determining that a vehicle is equipped with an 
EDR, particularly since such data do not provide information on pre-
crash events. Siemens VDO Automotive AG characterized the rule as 
``semi-compulsory.''
    We have carefully considered the arguments presented by the 
commenters for requiring the installation of EDRs in all subject 
vehicles.
    We are not yet persuaded that it is necessary or appropriate to 
mandate the installation of EDRs. We believe that the industry's 
voluntary development and installation of EDRs, combined with the 
standardization requirements in this rule, will be sufficient to meet 
the agency's and public's near term needs. Standardized EDR data from 
the growing population of vehicles with EDRs, collected and compositely 
analyzed, will enable the agency to investigate crashes more 
effectively and to analyze safety equipment performance, resulting in 
improved agency understanding of crash and injury causation. These data 
will also lay a foundation for advanced crash notification systems.
    Further, insofar as achieving those near term goals is concerned, 
adopting a rule mandating EDR installation would result in an 
unnecessary cost for automobile manufacturers and consumers. To 
operate, EDRs need a databus.\26\ Since less expensive vehicles are not 
equipped with a databus, a rule mandating EDR installation would 
require manufacturers to install a databus in those vehicles. While we 
are not presently compelling the installation of EDRs, it is our 
intention that their use continue to expand.
---------------------------------------------------------------------------

    \26\ The bus (connections between and within the central 
processing unit, memory, and peripherals) is used to carry data.
---------------------------------------------------------------------------

    As for the agency's longer terms goals related to EDRs, we expect 
the extent of installation in new vehicles to continue increasing and 
to reach approximately 85 percent by model year 2010. Based on 
currently available information, such as that obtained in connection 
with our NCAP program, the new vehicles lacking an EDR in that model 
year will be primarily those manufactured either in Germany or Korea. 
As Korea has expressed interest in the development of an EDR standard 
under the International Standards Organization, it appears that Korean 
built vehicles also might eventually be voluntarily equipped with EDRs.
    Further, we believe that allowing the current voluntary, gradualist 
approach to increased installation of EDRs to continue is more 
appropriate for meeting those longer term goals than mandating an 
acceleration of further increases in the extent of installation. We are 
aware that some consumers are concerned about the ownership and use of 
EDR data. The voluntary approach provides additional time for 
implementing measures concerning those concerns.
    We have considered the comments of Advocates and SISC, asking us to 
mandate EDRs so that it is possible to gather additional data for 
safety research. The agency seeks to gather EDR information in a 
readily usable manner to analyze crashes and the performance of safety 
equipment as composite information (i.e., to discover statistically 
significant trends). We

[[Page 51011]]

believe that the current level of EDR installation, combined with our 
standardization requirement, will yield data of statistical 
significance. The expected further increases in the extent of 
installation will improve the quality of our data still further. In 
light of our expected ability to meet these near term goals, we do not 
see the need to mandate EDR installation at this time.
    We will monitor future increases in the extent of installation of 
EDRs and revisit this issue if appropriate.
    We do not agree with the comments that our decision to adopt data 
standardization requirements without also mandating the installation of 
EDRs will induce manufacturers to remove EDRs from the vehicles in 
which they are currently installed or to drop plans for installing them 
in additional vehicles. The fact that approximately two-thirds of new 
vehicles are already equipped with EDRs is strong evidence of a 
significant incentive to install these devices. Further, as noted 
below, the data standardization requirements we are adopting in this 
final rule are less extensive and thus less costly that the ones we 
proposed in the NPRM. More specifically, we lowered the number of 
events and elements to be recorded. Based on our cost estimates 
(discussed below), we do not believe that adoption of our revised data 
standardization requirements will increase costs sufficiently to create 
a countervailing incentive for manufacturers to remove EDRs. We also 
note that consumer products, such as OnStar[supreg], incorporate EDRs 
into their services. The consumer appeal of these consumer products 
strengthens the existing incentive for manufacturers to install EDRs in 
their vehicles.\27\ In sum, we conclude that there are major benefits 
from the use of EDRs, but the marketplace appears to be adopting EDRs 
and we do not currently see a need to mandate their installation. The 
agency will monitor further progress in this area, and will be prepared 
to consider this question further if needed.
---------------------------------------------------------------------------

    \27\ In response to the concerns that the breadth of our 
proposed EDR definition could have the effect of requiring the 
installation of EDRs, we note that we have revised the definition of 
EDR, as discussed below, to exclude static freeze-frame data 
elements.
---------------------------------------------------------------------------

    We have also considered GM's comment urging us to ban EDR ``on/off 
switches'' and the comments of other commenters asking that we require 
them. This final rule concerns the standardization of EDR data elements 
and ensuring that downloaded EDR data are available to intended users. 
We did not propose either requiring or precluding on/off switches in 
the NPRM. We note, however that on/off switches could limit the 
benefits provided by EDRs. As with the issue of mandating installation 
of EDRs, we think it premature to compel action on this issue, and will 
continue to monitor and assess whether action is warranted in the 
future.
2. EDR Data Elements
a. Number and Types of Required Data Elements
    The NPRM provided a list of required data elements (a minimum set 
of elements required to be recorded if a vehicle is equipped with an 
EDR, regardless of whether those elements are presently recorded by the 
vehicle's EDR) and a list of ``if equipped'' elements (elements that 
would be required to be recorded only if the vehicle is equipped with 
the relevant safety system or sensing capability).
    NHTSA received several comments on the proposal's number of 
required data elements. Several manufacturers commented that the 
proposal's required number of data elements was excessive; however, 
manufacturers' comments differed as to which of the data elements 
should be deleted. Commenters representing highway safety advocacy 
groups suggested that the number of required elements is insufficient 
to meet NHTSA's stated goals of improving data compatibility, crash 
investigation, and safety. Some commenters suggested adding the VIN as 
a required data element.
    GM, DaimlerChrysler, IIHS, and Mitsubishi argued that the NPRM 
proposed an excessive number of data elements. GM provided a critique 
of the each of the data elements and recommended a different list of 
required data elements. GM's position was that the NPRM's data elements 
go beyond the minimum set of data elements needed by safety researchers 
and crash reconstructionists. GM argued that the number of required 
elements in the NPRM could compromise the ability of the vehicle's 
control modules to perform their primary function of deploying 
restraint systems. The number of required elements could also inhibit 
manufacturers from collecting other, more potentially useful data, to 
the extent that the required elements consume available processing 
capacity.\28\ IIHS made a similar comment, stating that the number of 
proposed data elements increases the burden on manufacturers and the 
incentive for manufacturers to delay or eliminate safety features. 
Mitsubishi commented that NHTSA should only require those data elements 
that are needed to capture crash data that would truly be useful in 
improving motor vehicle safety.
---------------------------------------------------------------------------

    \28\ Accordingly, GM, Daimler Chrysler, Ford, and Toyota 
recommended deletion of the following ``required'' data elements: 
(1) Engine RPM; (2) Longitudinal Acceleration (x-direction); (3) 
Multi-event Crash; and (4) Frontal Air Bag Deployment Level. The 
four automakers also recommended deletion of the following ``if 
equipped'' data elements: (1) ABS Activity; (2) Lateral Acceleration 
(y-direction); (3) Normal Acceleration (z-direction); (4) Occupant 
Size Classification; (5) Seat Position; (6) Steering Wheel Angle; 
(7) Stability Control; (8) Frontal Air Bag Suppression Switch 
Status; (9) Vehicle Roll Angle; (10) Disposal (second stage of a 
frontal air bag).
---------------------------------------------------------------------------

    Hyundai and Kia offered several comments regarding NHTSA's proposed 
data elements. First, they requested that, ``data capture be limited to 
events that trigger air bag deployment.'' Second, they commented that 
``engine RPM'' and ``engine throttle'' data serve the same purpose and 
requested that only one of those data elements be required. Third, 
Hyundai and Kia commented that the data elements ``Ignition cycle, 
crash'' and ``Ignition cycle, download'' should not be required; 
Hyundai's and Kia's position is that these data elements do not provide 
data about the crash event, and that these elements would require 
additional programming and memory. According to these companies, 
requiring these data elements would increase costs and necessary lead 
time.
    Delphi recommended that NHTSA limit ``the content of event records 
to those data that are of significant value to crash investigation and 
safety system performance analysis' in order to reduce the amount of 
memory that will be required. Delphi stated that each required 
parameter would consume memory for six instances of that parameter 
because of the need to hold and compare up to three events in temporary 
and permanent memory.
    Subaru and AIAM argued that the NPRM contained too many data 
elements, and each provided a recommendation for which data elements 
the final rule should require. Subaru recommended that NHTSA should re-
select and prioritize data elements in order to increase the 
feasibility of compliance with a final regulation. Specifically, Subaru 
recommended that NHTSA ``omit acceleration direction, tolerance range, 
and accuracy of G sensors from the requirement or allow significant 
additional lead time on a phase-in schedule.'' AIAM commented that to 
reduce the number of systems that would require a complete redesign of 
vehicle electrical architecture, the minimum data set should include 
only the following data elements: (1) Driver and front passenger belt 
use; (2) throttle

[[Page 51012]]

position; (3) brakes on/off; (4) ABS engaged/not engaged; (5) vehicle 
speed; (6) longitudinal and lateral vehicle acceleration; (7) delta-V; 
and (8) time of air bag deployment.
    In contrast to the commenters who suggested that the NPRM contains 
too many required data elements, Public Citizen and PCIAA encouraged 
NHTSA to require additional data elements. Public Citizen stated that 
to maximize the benefits of the EDR rule, NHTSA should standardize 
(i.e., require) a far more extensive list of EDR data elements. Public 
Citizen pointed to the Institute of Electrical and Electronics 
Engineers (IEEE) Project 1616 (``Motor Vehicle Event Data Recorders''), 
which includes 80 EDR data elements used by different groups. Public 
Citizen commented that NHTSA did not propose to require many of the 
``top ten'' data elements listed by the NHTSA-sponsored EDR Working 
Group.\29\ Public Citizen argued that standardizing EDR data elements 
would ensure compatibility of EDR data.
---------------------------------------------------------------------------

    \29\ We note that this group was a fact-finding group, and the 
findings were those of the group and not NHTSA's findings.
---------------------------------------------------------------------------

    PCIAA commented that the proposed rule focuses too much on 
restraint systems and not enough on systems to help the driver avoid 
collisions. PCIAA suggested that NHTSA should require data elements 
that would track driver inputs and the performance of the steering, 
suspension, or braking systems. According to PCIAA, the rule should 
include other equipment such as vehicle lighting or ``intelligent 
vehicle'' systems and should address equipment that is likely to be in 
used in the future, such as stability control systems, radar, cameras, 
and similar technology to monitor the driving environment.
    Nissan, Mr. Fink, Mr. Kast,\30\ Bendix, and AAA all suggested 
specific data elements they believe should be required. Several data 
elements that the commenters suggested we require were not proposed in 
the NPRM. Nissan suggested that the following elements be required: (1) 
Delta-V direction (lateral, longitudinal, vertical); (2) roll rate 
(roll acceleration); (3) yaw rate; (4) gear position; (5) traction 
control system status; (6) number of downloads after event; and (7) 
passenger air bag disable indicator status. Mr. Fink stated that the 
rule should require a standard data set, including ``vehicle speed, 
brake switch status, accelerator status, engine rpm, seat belt switch 
status and air bag deployment/belt pre-tensioner status.'' Mr. Kast 
commented that, based on his studies of EDR data, the following 
elements are necessary to evaluate the cause of a crash: (1) Status of 
dimmed headlights; (2) status of high beam; (3) status of indicator 
left; (4) status of indicator right; (5) status of any special signals; 
and (6) yaw angle or yaw angle velocity. Mr. Kast's rationale is that 
the status of the lighting equipment and turn signals are important for 
the evaluation of crashes that occur in the dark. Mr. Kast also 
emphasized the importance of knowing the yaw angle or yaw angle 
velocity in order to calculate the trajectory of the vehicle.
---------------------------------------------------------------------------

    \30\ Mr. Kast's comments were submitted independently and by 
Siemens VDO Automotive, AG. According to Mr. Kast, he is an 
``independent expert in the field of accident investigation and 
accident data recorders.''
---------------------------------------------------------------------------

    Bendix Commercial Vehicle Systems, L.L.C. commented that the 
following data elements should be included in the minimum requirements: 
(1) Transmission status (gear selection on automatic transmissions); 
(2) brake switch status; (3) accelerator (%); (4) engine speed (RPM); 
(5) date time; (6) engine hours; (7) odometer reading; (8) headlights 
on/off; (9) turn signal status; (10) cruise control (on/off); (11) ABS 
fault status; and (12) tire pressure (axle or each wheel or as 
regulated by NHTSA).
    AAA commented that rear seat air bags are being installed with 
increasing frequency and stated that NHTSA should consider requiring 
the recording of data elements associated with rear seat air bags in 
vehicles so equipped (e.g., rear seat occupant presence, size, seating 
position, and restraint use).
    SISC, Children's Hospital of Philadelphia, Delphi, and Public 
Citizen commented on the NPRM's categorization of data elements as 
``required for all vehicles equipped with an EDR'' (Table I) or 
``required for vehicles under specified conditions'' (Table II) and 
suggested that we change the categorization of certain data elements. 
SISC stated that NHTSA should mandate lateral acceleration as part of 
the required set of data elements. According to SISC, multi-axis 
accelerometers are becoming less expensive, and both longitudinal and 
lateral acceleration are essential to determining the true delta-V and 
the principal direction of force, which are critical elements of 
general crash investigation, biomechanics research, and the 
understanding of injury causation. Children's Hospital of Philadelphia 
made a similar comment, stating lateral acceleration should be a 
mandatory data element.
    Delphi suggested that data elements not normally part of the 
restraint control system should be moved from the table of data 
elements required if the vehicle is equipped with an EDR to the table 
(NPRM's Table II) of elements required under specific conditions (e.g., 
vehicle indicated speed, engine RPM, engine throttle, service brake 
status). According to Delphi, this would lower the cost of 
implementation for many manufacturers.
    On the other hand, Public Citizen argued for re-categorizing 
several data elements that the NPRM proposed to be recorded only under 
specified conditions (Table II) and instead require them (i.e., place 
them in Table I). Public Citizen believes that the final rule should 
require these data elements to be recorded (e.g., seat belt status for 
the front passenger). Public Citizen's rationale is that many of these 
elements only require additional sensing capabilities, which are fairly 
inexpensive in most cases.
    NTSB expressed concern that Table I and Table II will result in 
different data being available from different EDRs. It stated that the 
rule should require the same information from all EDRs to encourage 
uniformity of data and standardization of EDR usage. NTSB encouraged 
NHTSA to develop a comprehensive standardized list of data elements 
that would apply to all highway vehicles, including heavy vehicles.
    Several commenters, including Mr. Kast, Injury Sciences, Public 
Citizen, and EPIC, recommended requiring some type of date/time stamp 
and/or VIN information. Mr. Kast and Injury Sciences commented that a 
date/time stamp should be added to the required elements in order to 
correlate the recorded data with a crash event. Mr. Kast explained that 
the linkage is particularly important since low intensity accidents may 
be recorded. If this information is not required, Injury Sciences urged 
NHTSA to consider alternatives for linking data to a particular vehicle 
and accident. Public Citizen stated that a VIN data element would 
significantly increase the usefulness of EDR data by permitting 
crosschecks across various NHTSA databases. EPIC commented that the EDR 
should record the first eleven digits of the VIN, although ``the unique 
serial number portion of the VIN--a personal identifier--should not be 
collected.'' EPIC's rationale is that make, model, and manufacturing 
origin are important data for crash analysis.
    GM and Delphi raised cost issues pertaining to the data elements. 
GM requested that the final rule expressly state that the specified 
list of data elements is not intended to limit manufacturers' ability 
to voluntarily collect and record additional data elements. Delphi 
suggested that the

[[Page 51013]]

condition for an element to be required (Table II of the NPRM) be 
changed from ``vehicle is equipped,'' to ``data is available to the 
recording device.'' According to Delphi, this would lower the cost of 
implementation for many manufacturers.
    Siemens VDO Automotive AG and Bendix commented on the state of 
technology and our EDR proposal. Siemens VDO Automotive AG commented 
that the NPRM definitions for data elements should be modified (i.e., 
made more stringent) to reflect the state of technology already 
available and in use. Siemens predicted that the changes would not 
result in significantly higher costs because the standardization and 
adoption by all manufacturers would lower the costs of production. 
Bendix suggested that solid state digital storage media and non-
volatile storage devices could be used in conjunction with emerging 
technologies in the area of high-speed data links, which combine data, 
voice, and video data on a single communications link to record 
additional types of data.
    Nissan and Honda requested clarification on specific technical 
aspects of our proposal. Nissan stated that instead of recording the 
engine throttle, we should require recording the accelerator pedal 
operation. Additionally, Nissan suggested that the rule should permit 
two alternatives for determining the beginning of an event, as provided 
in SAE J1698, Vehicle Event Data Interface--Vehicular Output Data 
Definition. Nissan also sought clarification about the ``complete file 
recorded'' data element. Nissan questioned whether the ``Yes'' value 
indicates that the EDR functioned the whole time or whether the data 
set is complete (i.e., the EDR received good data from all systems). 
Honda sought clarification related to the data element for ``frontal 
air bag deployment level.'' Honda sought to confirm its understanding 
that this term means the percentage of maximum inflator output used for 
occupant restraint (i.e., inflator output excluding the output of the 
deployment for disposal, regardless of the delay timing of the second 
(disposal) stage deployment).
    TRW Automotive commented that the status of the anti-lock braking 
system (ABS) is not adequately indicated by the ``ABS Activity'' data 
element. TRW suggested that ``ABS Warning Lamp (On/Off)'' would provide 
a better indication of the status of the ABS system at the time a crash 
occurred. TRW Automotive commented that the data attributes for 
stability control systems should be modified because they do not 
clearly indicate the status. According to TRW, ``off'' should indicate 
that the driver has turned off the system, and an attribute ``Not 
Available'' should be added to indicate that the system is in a ``not 
available'' state.
    We indicated in the NPRM that it was not our intention to require 
manufacturers to install expensive technological hardware or software 
to meet our EDR standardization proposal. In the NPRM, we emphasized 
that vehicle manufactures have voluntarily made significant investments 
in EDRs and are already recording several data elements that suit our 
goals. The NPRM explained that our proposal sought to build upon the 
automotive industry's EDR accomplishments by standardizing the way data 
elements are captured and recorded. In other words, we considered our 
proposal to record the most important data elements relevant to crash 
reconstruction, the analysis of safety equipment performance, and ACN 
in light of the data already being processed by vehicles.
    We envisioned and it was our intent that the proposed EDR 
standardization requirements could be implemented by vehicle 
manufacturers at a minimal cost, since vehicle manufacturers had made 
EDR capability an additional function of a vehicle's air bag control 
system. We did not intend to require vehicle manufacturers to install 
equipment, such as additional accelerometers, to comply with the rule. 
(We estimated, for example, that an additional accelerometer could cost 
$20 per vehicle.)
    Our approach of standardizing the most important data elements at a 
minimal cost remains the same. However, after carefully considering the 
comments, we have re-evaluated the number and types of data elements 
that manufacturers should be required to standardize. We learned from 
the comments that the frequency, range, accuracy, and precision 
requirements (discussed subsequently) for many of the data elements we 
proposed would require an upgrade in sensors, microprocessors, and 
memory capability that would substantially add to the cost of complying 
with this rule. This was not our intention. We also learned that it is 
not current industry practice to record some of the data elements we 
proposed. In order to remain consistent with our approach of 
standardizing data at a minimal cost, we have revised the number of 
required data elements to reduce implementation cost and better reflect 
current industry practice.
    In revising the number and types of data elements to be recorded if 
a vehicle is equipped with an EDR (i.e., Table I), we deleted five 
items that we had proposed in the NPRM: ``longitudinal acceleration,'' 
``engine RPM,'' ``frontal air bag deployment level, driver,'' ``frontal 
air bag deployment level, right front passenger,'' and ``time from 
event 2 to 3.'' We added two items: ``time, maximum delta-V'' and 
``delta-V, longitudinal.'' \31\
---------------------------------------------------------------------------

    \31\ A discussion of our changes relating to the acceleration 
and delta-V data elements occurs in the next subsection, titled 
``The Acceleration and Delta-V Data Elements.''
---------------------------------------------------------------------------

    We deleted the ``engine RPM'' from Table I but added it to Table 
II. ``Engine RPM'' is somewhat related to ``accelerator pedal 
position.'' Accelerator pedal position reflects the driver's input to 
the engine. Engine RPM indicates the engine's response to that input. 
We believe that the two data elements are closely related, although 
distinct. We have reviewed many of GM's EDR crash data sets, and see 
little value in requiring ``engine RPM'' at this time. Moving this data 
element to Table II will reduce memory costs and the amount of data 
manipulation during pre-crash.
    After carefully considering the comments, we have also decided to 
remove ``frontal air bag deployment level, driver'' and ``frontal air 
bag deployment level, right front passenger'' from the list of required 
data elements (Table I). These elements would have indicated the 
deployment level of the driver's and right front seat passenger's air 
bag system. After further consideration, we believe that the same 
information we anticipated gathering from these deleted data elements 
can be ascertained using other data elements: ``frontal air bag 
deployment, time to deploy, in the case of a single stage air bag, or 
time to first stage deployment, in the case of a multi-stage air bag, 
driver'' (and the right front passenger equivalent) from Table I and 
``frontal air bag deployment time to nth stage, driver'' 
(and the right front passenger equivalent) from Table II.
    In revising the number and types of data elements to be recorded 
under specified conditions (Table II), we added four items that did not 
appear in the NPRM: ``delta-v, lateral,'' ``maximum delta-V, lateral,'' 
``time to maximum delta-V, lateral,'' and ``time to maximum, delta-V, 
resultant.'' \32\ Commenters had requested changes in the data elements 
for longitudinal acceleration and delta-V, and as noted elsewhere in 
this document, the agency has adopted a number of those changes

[[Page 51014]]

as part of this final rule. However, in order to fully implement those 
changes for the longitudinal direction, we believe it is necessary to 
also adopt data elements that constitute the lateral counterpart of the 
requested changes. This was done to provide standardized data elements 
that are consistent with those in Table I for longitudinal acceleration 
and delta-V. However, we have incorporated these additional data 
elements in Table II, rather than Table I.
---------------------------------------------------------------------------

    \32\ A discussion of the data elements related to acceleration 
and delta-V follows below in section titled ``The Acceleration and 
Delta-V Data Elements.''
---------------------------------------------------------------------------

    After considering the comments, we have decided to retain a number 
of the data elements that some manufacturers recommended that we 
delete, including ``occupant size classification'' and ``frontal air 
bag suppression switch status.'' Occupant size classification is 
important in determining whether the advanced restraint systems are 
working properly by drawing a comparison between the occupant and the 
safety system's classification. We believe that this is vital to that 
purpose of obtaining EDR data for the analysis of safety equipment 
performance. Frontal air bag on/off switch status is important in cases 
where the right front passenger air bag does not deploy. There is a 
possibility in some vehicles with no back seats that the air bag was 
turned-off at the time of the crash. It is critical that the EDR 
capture this evidence to enable an evaluation of whether advanced 
restraint systems functioned properly.
    We have also decided to retain ``ignition cycle, crash'' and 
``ignition cycle, download.'' These two data elements provide a method 
to identify whether the data stored in the EDR is related to a crash 
under investigation or to a previous crash.
    As indicated above, several commenters recommended recording other 
data elements that we did not propose to record (e.g., roll rate, yaw 
rate, gear position, number of downloads after event, passenger air bag 
disable indicator status, status of lamps and signals, engine hours, 
odometer reading, cruise control, ABS fault status, ``intelligent 
vehicle systems,'' steering input, and tire pressure). We have 
carefully considered these recommendations. We emphasize this final 
rule standardizes and requires (Table I) the most important data 
elements that are essential to crash reconstruction, the analysis of 
safety equipment performance, and ACN. We have decided not to require 
the recording of these additional data elements. We believe that 
recording these additional data elements, which are currently of lesser 
value for our stated purposes, would not only result in significantly 
higher costs but would also risk overburdening the microprocessing and 
memory capabilities of EDRs and increase potential record times. This 
increases the risk of system failure. We may revisit the distribution 
of data elements between Table I and Table II as technology advances, 
costs decrease, and the ability to record these data elements become 
less risky. We may also consider expanding Table II in the future as 
manufacturers expand the capability of EDRs and add additional sensors 
to motor vehicles that could be beneficial to motor vehicle safety.
    We have carefully considered comments from Mr. Kast, Injury 
Sciences, Public Citizen, and EPIC that we should include crash 
location, date/time stamp and VIN as data elements. We believe that the 
data elements related to crash location, date, time, and VIN are not 
essential to meet our goals of crash investigation and safety equipment 
performance. As we have stated earlier, we are currently standardizing 
only data elements that are important for composite analysis. We have a 
need to gather information about specific crashes only as it is related 
to general trends that we may discover with the information we gather. 
Therefore, we presently find it unnecessary to require manufacturers to 
collect or to standardize this type of data.
    After considering Public Citizen's comments, we disagree with the 
argument that the final rule must include all elements listed in the 
IEEE 1616 MVEDR Standard report and the ``top ten'' items presented in 
the NHTSA-sponsored EDR working group report. The IEEE 1616 report, 
which lists 80 data elements, is a compilation of the data elements 
that are available/recordable at present, or expected to be in the 
future, for various vehicles. In other words, the data elements listed 
in the IEEE 1616 report are a compilation of all available data 
elements (i.e., a ``data dictionary''), and not a recommended set of 
data elements. We do not believe it would be appropriate at this time 
to require automobile manufacturers to record all of the data elements 
contained in the IEEE 1616 report. Doing so would substantially extend 
the number of standardized data elements, resulting in redundancy and 
the standardization of many data elements that are presently unrelated 
to the purposes of this rulemaking.
    Public Citizen also contends that we have not included many of the 
data elements listed in the ``top ten'' list found in the NHTSA-
sponsored EDR working group report. As we stated above, this group was 
sponsored by NHTSA; however, we have never adopted its findings as our 
own. However, we note that the final rule does include standardization 
protocols for many of the same data elements that are listed in the 
``top ten'' list, including longitudinal/lateral acceleration, seat 
belt status, pre-crash data (e.g., steering wheel angle, brake use, 
vehicle speed), vehicle roll angle, ABS, stability control, and air bag 
data.
    We have considered Delphi's recommendation that data elements not 
pertinent to restraint control and/or crash reconstruction should be 
moved to Table II. Our NPRM was based on this premise; that is, Table I 
contains the data elements critical to crash reconstruction, advanced 
restraint operation, and enabling ACN. We continue to support our 
approach for Table I data elements. However, we have modified Table I 
slightly by: (1) Moving the ``engine RPM'' data element to Table II, 
because it can generally be inferred from accelerator pedal position; 
(2) substituting delta-V-related crash severity measurements for 
acceleration measurements to reduce complication and cost of EDRs, and 
(3) dropping those data elements related to a third event, because we 
believe two events will capture most crashes with multiple, non-trivial 
events.
    NTSB expressed its desire for recording accelerator/brake pedal 
positions in certain special crashes. The revised Table I retains both 
of these data elements.
    We have also considered Public Citizen's arguments that several 
data elements currently listed in Table II could be moved to Table I 
(required) for minimal to no cost (e.g., safety belt status, front 
passenger). The costs associated with placing particular data elements 
in Table I is not the sole factor in determining whether to include 
that data element in the core set listed in Table I. To minimize the 
risk of data loss, we must also consider the current capabilities of 
microprocessors to process the information and the availability of 
memory storage capacity. The longer or larger the data file, the more 
complicated it is to record it successfully during a crash. We believe 
our Table I required list and our Table II (standard formats for data 
elements recorded by manufacturers) provide a reasonable balance of 
these concerns and priorities. We may reevaluate the number and types 
of data elements in the future as EDRs, memory, and microprocessing 
continue to develop. In the meantime, we believe that it is appropriate 
to keep ``safety belt status, right front passenger'' and other similar

[[Page 51015]]

data elements in Table II. We emphasize that our final rule requires a 
minimum set of data, and manufacturers may and most likely will exceed 
this minimum data set, incorporating data elements listed in Table II 
and event data elements we have not listed in this final rule.
    We note that the data elements in Table II must be standardized if 
recorded. Therefore, we believe that manufacturers that are currently 
recording these data elements will be able to standardize at a minimal 
cost.
    We have considered NTSB's comment encouraging uniformity in the 
number and types of data elements recorded for all EDRs. As noted 
above, we believe this final rule standardizes a core set of data 
elements that will be useful for crash reconstruction, the analysis of 
safety equipment performance, and the development of ACN. Table I does 
standardize a core set of data elements among all vehicles equipped 
with an EDR. However, we recognize that vehicle manufacturers are in 
different stages of technological development with their EDRs. Some 
manufacturers have made greater strides in the development of EDRs and 
the number of recorded data elements while others have been slower to 
evolve. We developed Table II to standardize data elements that are 
currently recorded by some manufacturers, but not others.
    NTSB commented that they were concerned that our approach of a 
minimum data set (Table I) combined with an optional data set (Table 
II) would result in different vehicles recording different data 
elements. This regulation establishes a minimum data set for vehicles 
that are equipped with an EDR. Manufacturers are permitted to record 
other additional data elements, as they believe fit the needs of their 
vehicles and equipment installed on their vehicles. We have taken an 
approach that will: (1) Ensure that all vehicles equipped with an EDR 
will have a minimum set of data, (2) provide standardization for 
additional priority data elements, and (3) allow manufacturers to 
obtain other additional data as they deem appropriate to meet their 
needs. As EDRs evolve, NHTSA may reevaluate this approach in future 
rulemakings.
    We also want to provide manufacturers with the flexibility to 
improve their EDR designs and record a diverse group of data elements 
so that we may continue to study the usefulness of various data 
elements in terms of safety. We view EDRs as a new technology that has 
not seen much maturation outside of its initial inception as part of 
the air bag module. This rulemaking, we believe, is a positive step 
toward guiding the development of EDR technology for vehicle safety 
purposes by both requiring a standardized set of data elements that we 
believe will be useful while at the same time providing manufacturers 
with the ability to continue to evolve the EDR.
    Likewise, we do not agree with IIHS that the number of data 
elements we have chosen to standardize will provide incentive for 
manufacturers to delay or eliminate safety features. Our cost estimates 
indicate that our standardization requirements, as revised in this 
final rule, will not result in significant costs to manufacturers. As 
stated above, we have narrowed the recording requirements for EDRs. 
Also, we believe that our decision to standardize a core set of data 
elements and requiring standardization of data elements ``if recorded'' 
will allow flexibility for manufacturers to research and develop EDRs.
    We have considered Hyundai's and Kia's argument that we should 
limit the recording of data to events that trigger air bag deployment. 
We do not believe that limiting our data to events that trigger air bag 
deployment would be sufficient for our purposes. We want to know about 
events that should have deployed air bags, but did not do so, 
indicating the possible existence of a defect. Further, we seek to 
gather data not only to analyze the performance of air bags, but also 
to analyze the performance of other safety equipment, such as seat 
belts. We also seek to gather data helpful for crash reconstruction. We 
believe that this data can be standardized and recorded without 
significant cost. Further, we anticipate that development of e-911 and 
ACN systems may lead vehicle manufacturers to incorporate additional 
elements besides air bag deployment; such elements may provide 
information to EMS regarding other crash modes, such as side impact and 
rollover, as sensor technologies advance and their costs decline.
    We do not agree with AIAM that our final rule will require a 
complete redesign of vehicle electrical architecture if we do not 
reduce the minimum data set to the eight elements it proposes. As 
discussed in the costs section, we anticipate negligible redesign to 
the electrical architecture of vehicles as a result of our final rule. 
Additionally, we note that our new Table II is similar to AIAM's 
recommendation. Our Table II includes ABS engaged/not engaged and right 
front passenger belt use.
    Nissan requested clarification about the ``complete file recorded'' 
data element, asking whether the ``yes'' value indicates that the EDR 
functioned the whole time or whether the data set is complete. A 
complete record is a record that ends normally, regardless of the 
amount of data. An incomplete record is one that ends abnormally. For 
example, a complete value with ``yes'' indication would include a 
scenario where all data elements were captured successfully and 
recorded to memory or where some elements were not captured because of 
device failure but the full record was recorded to memory. Examples of 
when there is an incomplete record is where all data was captured 
successfully, but the record function interrupted and the file is 
incomplete, or in the case of a power or system failure, there is no 
data captured, so there is no value.
    TRW commented that the data element ``ABS Activity'' does not 
adequately indicate the status of the ABS system. In the NPRM, we 
intended the word ``status'' to mean that the ABS was actively 
controlling the brake forces, not whether the system status was 
operational. We would expect ``on'' to mean that the vehicle's ABS was 
actively controlling the vehicle brakes. Conversely, we would expect 
``off'' to be used at all other times. For example, if a person is 
stopping and presses the brakes moderately in normal driving 
conditions, then we would expect the service brake operation to 
indicate ``on.'' If driver uses hard braking, activating the ABS, then 
the ABS activity would indicate ``on'' for that time period. The 
``service brake'' data element would continue to read ``on'' during 
periods of ABS activity.
    TRW also commented that the data attributes for stability control 
systems should be modified because they do not clearly indicate the 
status. We proposed three states for stability control: ``on,'' 
``off,'' and ``engaged.'' ``On'' and ``off'' are intended to be status 
of the vehicle's stability control on/off switch. We intend ``engaged'' 
to be used when the stability control is actively controlling the 
vehicle. Some vehicles do not have on/off switches for stability 
control, and the systems remains ``on.'' In such a case, the indicator 
would read, depending on the circumstances, either ``yes'' or 
``engaged.''
    We also made a modification to the ``condition or requirement'' 
provision for most of the data elements in Table II. In the NPRM, we 
used the phrase ``if equipped.'' We proposed the phrase ``if equipped'' 
because we envisioned requiring manufacturers to record the data 
elements in Table II if the vehicle is equipped with the relevant 
safety

[[Page 51016]]

system or sensing capability. In the final rule, the condition or 
requirement for most data elements in Table II will be ``if recorded.'' 
By using ``if recorded'' we mean that manufacturers are required to 
comply with Table II if the data element is recorded in non-volatile 
memory for the purpose of subsequent downloading. We made this 
modification so that the final rule better reflects current industry 
practices. Some data elements may only be recorded in volatile memory 
(for applications such as air bag deployment) and not non-volatile 
memory for the purpose of subsequent downloading. Our proposal would 
have required manufacturers to record in non-volatile memory certain 
data elements, such as acceleration.
    We also believe that the change effectuates our goal of providing 
standardization for the data elements listed in Table II without 
substantial cost or risking EDR malfunction. We agree with the 
commenters that recording these data elements, such as acceleration, at 
the frequency and intervals we proposed, would require additional 
memory--adding to the cost of implementation. Recording these data 
elements in non-volatile memory would have also increased the risk of 
not capturing a complete crash record. A more complete discussion of 
the risks associated with recording large crash records is discussed 
below.
b. The ``Acceleration'' and ``Delta-V'' Data Elements
    In the NPRM, we proposed that Table I (the minimum data set) 
include the crash severity data elements ``longitudinal acceleration'' 
and ``maximum delta-V.'' We selected longitudinal acceleration to 
provide crash severity information. Longitudinal acceleration is a 
common data element collected in engineering studies and crash tests to 
determine crash severity and the shape of the crash pulse in frontal 
and rear crashes. It also provides information regarding the maximum 
acceleration level. Therefore, we believed that it was appropriate to 
standardize longitudinal acceleration.
    We also proposed to include maximum delta-V in the minimum data 
set. We proposed to include maximum delta-V in the minimum data set 
because it quantifies the severity of the crash in the vehicle's 
memory. We had proposed that the absolute value of maximum delta-V be 
used, if the vehicle experienced a second crash, to determine whether 
the data in the EDR's memory should be replaced with the subsequent (or 
second) crash information. We proposed that only subsequent crashes 
with higher maximum delta-V must be recorded in the vehicle's memory.
    GM, Daimler Chrysler, Ford, Honda, and Toyota specifically 
requested that we replace the longitudinal acceleration and lateral 
acceleration data elements, as proposed in the NPRM, with longitudinal/
lateral delta-V elements. All suggested that delta-V is a better 
indicator of crash severity than acceleration. They stated that while 
longitudinal/lateral acceleration is currently recorded by some 
manufacturers, acceleration data is not currently used or needed for 
safety-related crash analysis and reconstruction purposes. The data is 
intended for internal use, specifically to understand deployment 
algorithms. DaimlerChrysler explained that because of this very 
specific use of acceleration data, the time duration recorded was never 
intended to capture a complete crash and is usually too volatile for 
use in crash investigation. GM made a similar comment, stating that 
delta-V is preferred over acceleration in analyzing crash 
reconstruction because acceleration data, even after filtering, is 
typically too sporadic.\33\ Accordingly, the manufacturers stated that 
accident reconstructionists usually use delta-V instead of acceleration 
data.
---------------------------------------------------------------------------

    \33\ Recording intervals were suggested for the proposed delta-V 
recording element and are discussed below.
---------------------------------------------------------------------------

    Honda also recommended replacing the proposed elements for 
longitudinal, lateral, and normal accelerations with delta-V, coupled 
with the angle or direction of delta-V, to improve the overall 
understanding of a crash event.\34\
---------------------------------------------------------------------------

    \34\ Specifically, Honda recommended changing the definition of 
delta-V, replacing it with the following language:
    Delta-V means, for vehicles with only longitudinal acceleration 
measurement capability, the change in velocity of the vehicle along 
the longitudinal axis, and for vehicles with longitudinal, lateral 
and/or normal acceleration measurement capability, the magnitude and 
direction of the change in velocity of the resultant of the 
longitudinal, lateral and/or normal vehicle velocity time-histories, 
within the time interval starting from the time zero and ending 500 
ms after time zero.
---------------------------------------------------------------------------

    Hyundai and Kia suggested that the ``lateral acceleration'' and 
``normal acceleration'' data elements should not be required even if 
the vehicle is equipped with sensors. Hyundai and Kia stated that their 
tests have shown that the data acquired may be misleading due to 
external noise transmitted from body structure damage.
    After reviewing the comments, we have decided to adopt modified 
requirements for the collection and standardization of data associated 
with the acceleration and delta-V data elements. In the final rule, the 
acceleration data elements (longitudinal, lateral, and normal) will 
appear in Table II. In other words, the final rule will standardize 
acceleration data elements if manufacturers are recording the 
acceleration data elements. In lieu of longitudinal acceleration, the 
final rule focuses on recording delta-V as the crash severity measure.
    We have modified data elements relating to delta-V. In the final 
rule, Table I includes the data element ``delta-V, longitudinal,'' 
``maximum delta-V, longitudinal,'' and ``time, maximum delta-V.'' 
Delta-V longitudinal will provide for the tracking of longitudinal 
delta-V time series data, replacing our proposal to record the 
longitudinal acceleration time series. We are also adding a new data 
element to track the time associated with the maximum longitudinal 
delta-V.\35\ We are focusing on delta-V, modifying the final rule to 
enhance the standardization of delta-V data elements while also 
providing for the standardization of acceleration data if manufacturers 
record acceleration (now in Table II). We believe that delta-V will be 
sufficient to meet our purposes of analyzing safety equipment 
performance, aiding in crash reconstruction, and enabling ACN, while 
remaining sensitive to costs, the risk of data loss associated with 
writing large amounts of data to memory, and the problems associated 
with external noise transmitted from body structure damage.
---------------------------------------------------------------------------

    \35\ Our decision to record delta-V instead of acceleration 
resulted in the addition of five new definitions in the regulatory 
text: (1) Delta-V longitudinal, (2) maximum delta-V, longitudinal, 
(3) delta-V, lateral, (4) maximum delta-V lateral, and (5) time, 
maximum delta-V, lateral.
---------------------------------------------------------------------------

    We believe that delta-V is sufficient for our objectives. NHTSA has 
used delta-V as a measure of crash severity for many years. Delta-V is 
considered an essential part of crash investigation. For several 
decades, NHTSA's crash investigation teams have gathered information to 
estimate delta-V using computer programs. The EDR data will assist 
these researchers because they will be able to obtain a direct measure 
of delta-V.
    There are significant cost differences between delta-V and 
acceleration, notwithstanding that both of these time series 
measurements are typically based on accelerometer measurements. In 
current practice, acceleration time series data are collected every 2 
milliseconds for some EDRs while delta-V time series data are collected 
every 10 milliseconds

[[Page 51017]]

in others. Therefore, comparing these two practices, accelerometer data 
generates 5 times the volume of data. If we were to require, as 
proposed, longitudinal acceleration, it would be necessary to capture 
and record these data, increasing the cost. This increased cost is due 
to the increased size in the microprocessor, random access memory (RAM) 
and electrically erasable read only memory (EEROM) that would be needed 
to capture and record the volume of data produced by the longitudinal 
acceleration data element.
    In addition to cost, we have considered the comments that address 
the risk of data loss associated with recording a larger file (i.e., 
more data elements or data elements producing larger volumes of data, 
such as longitudinal acceleration). In explaining the risk of data 
loss, we first explain how the EDR records data. An EDR must 
continuously capture pre-crash data, and it must also capture crash 
data to determine if the trigger threshold has been met. If we required 
acceleration data, EDRs would be required to capture up to 150 
milliseconds of data, which equates to 76 data points. However, if we 
required delta-V data, EDRs would only need to collect about 26 data 
points, which would correspondingly reduce the amount of data to 
capture this element. Once the threshold has been met or exceeded, the 
remainder of the data set must be captured and then recorded. The 
actual recording operation takes place after the crash event. Severe 
crashes often interrupt (or destroy) the normal operation of the 
vehicle's electrical system. Interruption of the vehicle's electrical 
system may compromise the ability of the EDR to complete capturing and 
then record data. In the state of current technology, there is a much 
better chance of capturing and recording a complete file that is 
smaller rather than larger. Accordingly, we believe it is desirable to 
keep the file size (i.e., data elements/volume of data) to a minimum. 
As the state of technology improves and the cost of microprocessing and 
memory declines, we foresee the risk of data loss to pose less of a 
concern.
    In deciding to include delta-V in the minimum data set (Table I), 
we also considered the location of accelerometers. If the accelerometer 
is located in an area that has some small local movements (often called 
ringing) as a result of the crash, its acceleration profile will not 
match that of a rigidly attached accelerometer, producing different 
maximum deceleration measurements that would not be usable to make 
assessments for a vehicle's frontal crash stiffness--one of the 
measurements we were considering when we proposed acceleration as a 
required element. Our research indicates that, while acceleration 
profiles are not in good agreement between the EDR's accelerometer and 
a reference accelerometer, the delta-V measurements in such conditions 
are reliable.\36\
---------------------------------------------------------------------------

    \36\ ``Performance of Selected Event Data Recorders,'' Aloke 
Prasad, NHTSA 2001, available at http://ww-nrd.nhtsa.dot.gov/pdf/nrd-10/EDR/EDR-round-robin-Report.pdf.
---------------------------------------------------------------------------

    We have considered the comments asking us to include all directions 
of acceleration (x, y, and z) in our minimum data set. We believe that 
such information would be informative; however, we must balance our 
need against the cost and increased complications with expanding the 
minimum data set, as discussed above.
    The final rule does provide standardization protocols for the 
acceleration data elements if the manufacturer records them (Table II). 
Our decision to move longitudinal acceleration from Table I to Table 
II, as discussed above, rests on our belief that delta-V is sufficient 
for our present purposes, especially in light of the costs and risk of 
data loss that we face if we were to require the recording of 
longitudinal acceleration. We decided to retain lateral acceleration as 
a data element to be standardized if recorded in recognition that it is 
a data element that can provide useful information for crash 
reconstruction. We also expect lateral acceleration to become more 
useful as our proposed upgrade to FMVSS No. 214 evolves. Moreover, 
costs and the risk of data loss pose less of a concern on those 
manufacturers that have invested in their EDR programs to the point 
where they are recording longitudinal and lateral acceleration. We 
expect costs associated with merely standardizing the format of this 
data already recorded to be minimal.
c. Multiple-Event Crashes and the ``Multiple-Event'' Data Element
    In the NPRM, we proposed that the number of crash events be 
recorded as a data element, which is listed in Table 1 of the NPRM as 
``Multi-event, number of events (1, 2, 3).'' The proposed data element 
records the number of crash events (up to three events), with a maximum 
of 5 seconds as the proposed gap between connected events of a crash. 
Industry commenters disagreed with the NPRM's requirement to record up 
to three events--mostly because they believe such a requirement is 
technologically complex.
    According to GM, NHTSA's proposal did not provide a comprehensive 
and objective regulatory requirement with respect to multiple-events. 
GM stated that the final rule should not require EDRs to record data 
for multiple-impact crashes, but instead should only focus on single-
impact events. GM also argued that a regulatory requirement that 
focuses on recording single events would achieve NHTSA's regulatory 
objectives because most crashes involve single events. GM stated its 
belief that the multiple-event recording requirement is excessive in 
part because of the amount of buffering and data processing required to 
meet a regulatory requirement to record multiple-events could 
compromise the primary purpose of the module to properly deploy 
restraint systems and prevent crash injuries. GM urged modification of 
the NPRM's definition of ``event'' to reflect this change.
    Honda commented that the final rule should clarify an inconsistency 
in the NPRM related to recording of events in multi-event crashes where 
the air bag deploys. Honda stated that the NPRM provides that in a 
situation where the time after a ``trigger threshold event'' is less 
than 500 ms, subsequent ``event'' data would not be captured and 
recorded in a multi-event crash, even if there is air bag deployment 
(see definitions in Sec. 563.5). According to Honda, this conflicts 
with the intent of the data capture provisions in Sec. 563.9(d) of the 
NPRM. Therefore, Honda recommended that the final rule require, 
regardless of the time and/or recording status of any ``trigger event'' 
as defined in Sec. 563.5, that when air bags deploy, the ``event'' data 
should be recorded. Honda's rationale is that such a requirement will 
help ensure that EDR data will provide a better understanding of the 
circumstances of crashes that are severe enough to deploy an air bag.
    Nissan commented that the three-event requirement is unnecessary 
and would be expensive and technologically complex to implement. Nissan 
suggested that the elements related to the three-event requirement be 
dropped. However, if that requirement is retained, Nissan stated that 
NHTSA would need to clarify what constitutes a separate event and what 
combinations of events need to be recorded.
    AIAM commented that the recording of three events in a multi-event 
crash is not current industry practice. Instead that organization 
suggested that all recording stop after an event resulting in an air 
bag deployment. According to AIAM, recording three events, as specified 
in the NPRM, would be ``a major task'' and would require

[[Page 51018]]

additional memory and development of new software algorithms.
    Hyundai and Kia expressed concern that the accuracy of acceleration 
data captured after the first event is uncertain (if a multi-event 
crash involves two or more events in the same direction) because of the 
technical limitations of acceleration sensors currently available in 
the market for air bag systems. Based on this uncertainty, Hyundai and 
Kia recommended that we not require accurate acceleration data from an 
event that occurs after the air bag is deployed if this event occurs in 
the same direction as the previous event. Additionally, Hyundai and Kia 
suggested that ``recording time of longitudinal acceleration * * * be 
reduced from `-0.1 to 0.5 seconds' to `-0.1 to 0.3' seconds,'' arguing 
that this change would prevent recording overlap with other events and 
would reduce the implementation cost and time.
    Advocates supported the NPRM's proposal to record multi-event 
crashes, capturing up to three events, because of the high percentage 
of multiple-event crashes. However, Advocates asked us to reconsider 
whether a five second interval from the first triggering event would be 
sufficient to capture all or most of multi-event crashes.
    In light of the comments submitted on the multiple-event recording 
data element, we have decided to reduce the number of events to be 
recorded in a crash from three to two. We also decided to change the 
logic for capturing up to 2 events by limiting the capture to a single 
event in the event of a crash where an inflatable restraint is 
deployed. As a result, we have modified the data element to reflect up 
to 2 events in a crash, dropped the data element that recorded the time 
associated with event 3, and retained the data element that records the 
time between event 1 and event 2.
    We believe that reducing the multiple-event recording requirement 
to two events is appropriate considering the number of crashes that 
occur with two events or less. We believe the revision will also 
alleviate the additional cost and complications associated with 
recording up to 3 events. The following discussion explains our 
approach and rationale in further detail.
    Because we have, in effect, redefined an ``event'' as a change in 
delta-V that equals or exceeds 8 km/h (5 mph) in a 150 ms period, we 
needed to update our analysis in terms of what events are considered to 
be non-trivial, as would justify capture and recording by the EDR 
(i.e., events meeting the trigger threshold). In the NPRM, we proposed 
that EDRs must be capable of recording up to three events. In light of 
these changes, the agency re-examined the issue of multi-event 
recording in developing this final rule.
    NHTSA conducted an analysis using 2002 and 2003 National Automotive 
Sampling System/Crashworthiness Data System (NASS/CDS) data to 
determine the distribution of vehicles in multi-event crashes. This 
analysis provides a weighted annual estimate of the number of vehicles 
in crashes involving multiple events. The data from these two years 
reveal that approximately 3.2 million light vehicles in the United 
States were towed each year. Of these vehicles, about 2.25 million are 
involved in single-event crashes, and 0.85 million are involved in 
multi-event crashes. (The remaining 0.1 million had no event, suffered 
damage resulting strictly from rollover, or experienced some other non-
collision event.)
    Our analysis revealed that delta-V data are missing for at least 
one event for many of the 0.85 million vehicles involved in multi-event 
crashes. To avoid underestimating the frequency of vehicles involved in 
multi-event crashes, the analysis accounted for unknown delta-V data by 
adjusting the raw weighted estimate by the ratio of the number of 
relevant crashes to the number of crashes without any missing delta-V 
data. We assumed that the vehicles in multi-event crashes with unknown 
delta-V event data have crash severities similar to those in known 
delta-V crash events. Of the 0.85 million vehicles in multi-event 
crashes annually, 175,000 vehicles have delta-V data for all events, 
while the remaining 684,000 vehicles have at least one event with 
missing delta-V data. The total annual estimate of vehicles in multi-
event crashes where at least two of the events have non-trivial delta-
Vs is 587,000. The other 2.61 million vehicles were involved in crashes 
that had no more than one non-trivial impact.
    We have further estimated the distribution of vehicles experiencing 
exactly two non-trivial events, as compared to those experiencing three 
or more non-trivial events. (Again, this analysis uses the 
distributions established from the vehicles with known delta-V data to 
forecast the annual estimate.) Our analysis indicates that 
approximately 580,000 vehicles per year are involved in multi-event 
crashes with exactly two non-trivial events. The annual estimate of 
vehicles involved in crashes with three or more events is 6,000.
    In the final rule, we have also made a change in the data capture 
and recording strategy, and further allowed an exception to the 
multiple-event requirement. For each crash that has an event that 
exceeds the trigger threshold, the EDR records data, replacing data 
from the previously recorded event(s), up to two events. Typically, up 
to two events will be recorded. In those crashes where an air bag is 
deployed during one of the two events of the crash, only the event 
associated with the air bag deployment must be recorded. This exception 
is intended to ensure that a vehicle's microprocessors do not become 
overburdened during the critical period when the vehicle is deciding 
whether to deploy the air bag. (We note that while not required to do 
so, an EDR may capture multi-event data during a crash that involves an 
air bag deployment.)
    This exception in the capture/recording strategy may reduce the 
number of multi-event recordings (i.e., by the number associated with 
air bag deployments). Our analysis indicates that about 58 percent of 
the time when a vehicle is involved in exactly two non-trivial events, 
the air bags are not involved. (The ratio is about the same for 
vehicles experiencing one non-trivial event, and it is somewhat lower 
for vehicles experiencing three or more non-trivial events.) This 
estimate is based on frontal air bag deployment data. Factoring in 
these vehicles, the annual estimate of vehicles involved in crashes 
with two or more non-trivial events for which the EDR would need to 
capture and record data is reduced under the final rule, taking into 
account the air bag deployment crashes. We estimate that annually, 
about 340,000 vehicles would be involved in recordable non-air bag-
deployment crashes with two or more non-trivial events.
    For these reasons, NHTSA has decided to maintain the multi-event 
recording requirement in the final rule, but to reduce the number of 
events from three to two.
    Our modification from recording three events to two events will 
significantly reduce the amount of memory required, thereby addressing 
commenters concerns about memory and the multiple-event recording 
requirement. With regard to Hyundai's and Kia's concerns about the 
accuracy of acceleration and recording time, we believe that this issue 
is no longer relevant since we are no longer including acceleration in 
the minimum set of required data elements.
    In response to the comments asking us to clarify the multiple-event 
requirements, we will briefly discuss multiple-event recording. An 
event is defined as an impact or other physical

[[Page 51019]]

occurrence that meets the trigger threshold--5 mph (8 kph) delta-V 
within a 0.150-second period. When this occurs, the pre-crash data are 
frozen and the crash data are collected from time zero to 0.3 seconds.
    If the first event is the deployment of an inflatable restraint, 
these data are recorded to memory and the file is locked. No further 
analyses (i.e., looking for subsequent triggers) or recording occurs. 
If there is no inflatable restraint deployment during the first event, 
the data are captured and stored in a similar manner. There are several 
possibilities that could occur after this event
    First, no subsequent event occurs. In this case, the first event 
ends at 300 ms after time zero.
    Second, a subsequent event occurs without an air bag deployment. In 
this case, the first event ends at 300 ms and within 5 seconds of time 
zero (event 1) another event is detected. These data are then captured 
and recorded in a separate file, resulting in a two-event recording.
    Third, a subsequent event occurs where the second trigger is 
detected during the first event, that is, during the 300 ms data 
collection period of the first event. It is possible that a second 
impact in a multi-impact crash could occur while the first event is 
still being captured and recorded. In this case, the time between 
events could be less than 300 ms. This could occur in cases where the 
first event triggered quickly, such that the delta-V threshold (5 mph) 
was exceeded in just a few milliseconds, but it is also possible that 
it could occur anytime a subsequent time zero is detected before the 
end of the first event. In these cases, the second event would start 
the detection of the second trigger. It is the agency's intent that, in 
these cases, the EDR capture separate events and not different portions 
of a single event. Therefore, a method is needed to establish the end 
of the first event, so the agency has turned to SAE J1698-1, Vehicle 
Event Data Interface--Output Data Definition (March 2005), in resolving 
this issue. SAE 1698-1 defines the end of an event as the moment at 
which the cumulative delta-V within a 20 ms time period drops to 0.8 
km/h (0.5 mph) or less. Thus, in this special case, the EDR would not 
start looking for a new trigger threshold until the first event has 
ended. The pre-crash data could be the same for both events.
    Fourth, a subsequent event occurs with air bag deployment in cases 
where there is a pre-event (meets trigger threshold of delta-V greater 
than 5 mph) without an air bag deployment. The file associated with the 
air bag must be recorded and locked. The pre-air bag event may be 
recorded, but it is not necessary. We do not want the pre-crash event 
to affect the decision-making of the microprocessor, which has the 
primary function of analyzing the crash and properly deploying the air 
bags.
d. Sampling Rates and Recording Intervals for Required Data Elements
    The NPRM specified sampling rates and recording intervals for data 
elements in order to standardize EDR data across the entire spectrum of 
new makes and models of light vehicles. NHTSA received comments ranging 
from general concerns about the frequency of the rates and intervals to 
detailed comments concerning sampling rates and recording intervals.
    GM, Ford, DaimlerChrysler, and Toyota commented that the sampling 
rates and durations proposed in the NPRM are excessive in that the 
large number of data elements and prolonged recording time at a very 
high frequency rate will require memory storage capacity 5-10 times 
greater than the current memory capacity provided by manufacturers that 
have installed EDRs. These manufacturers further commented that 
recording data in the manner specified in the NPRM will increase the 
workload for the processor, which would most likely require an upgrade 
for the microprocessor. These system upgrades would add to the cost of 
complying with the data requirements.\37\
---------------------------------------------------------------------------

    \37\ The commenters did not provide a specific cost estimate.
---------------------------------------------------------------------------

    To address the memory storage capacity and microprocessor issues, 
GM, Ford, DaimlerChrysler, and Toyota recommended deleting several 
elements, as mentioned above, and provided alternative recording 
intervals/times and sample rates for specific data elements. In some 
instances, the alternative recording intervals/times and sample rates 
were suggested because their field experience shows no benefit to an 
accident reconstructionist for the additional recording time.\38\
---------------------------------------------------------------------------

    \38\ These elements include ``speed, vehicle indicated,'' 
``engine throttle, % full,'' and ``service brake''. For these 
elements, Ford, DaimlerChrysler, and Toyota recommended reducing the 
recording interval/time by three seconds from ``(-) 8 to 0 sec'' (as 
proposed in the NPRM) to ``(-5) to 0 sec.'' GM proposed ``-2.5 to 0. 
5 sec.'' All four commenters recommended reducing the sample rate 
per second from ``2'' (as proposed in the NPRM) to ``1'' for these 
data elements.
---------------------------------------------------------------------------

    Other commenters suggested technical changes to the recording 
times/intervals and sample rates for other reasons. Hyundai and Kia 
requested that NHTSA perform a cost/benefit analysis for the data 
elements with recording intervals from 150 ms to 500 ms after an event. 
They see little degradation in the quality of data captured with the 
shorter time period while the costs of implementation would be 
considerably higher with the 500 ms requirement.
    Delphi recommended that NHTSA change the recording period for all 
acceleration data from 500 ms after an event to only 200 ms after an 
event. Delphi recommended that NHTSA make the other related changes 
necessary to reflect this change in recording period (e.g., allowing 
the recorder to retrigger after 200 ms instead of after 500 ms).
    Honda recommended changing the ``vehicle roll angle'' measurement 
time interval from ``-1.0 to 6.0 sec'' to ``-1.0 to 0.5 sec.'' Honda's 
rationale is that because the proposed 563.9(a) specified that the EDR 
must collect data for an event starting at time zero and ending 500 ms 
later, the interval for vehicle roll angle must be adjusted to the 
required measurement time of 0.5 sec of a multi-crash event. Honda also 
stated that the time of air bag deployment should be recorded during 0 
to 500 ms to adjust to the acceleration measurement time. Honda 
commented that an air bag deployment event cannot be recorded 
separately while the acceleration after a trigger event is being 
recorded.
    While many automakers advocated for reducing the period of post-
crash recording, some advocacy groups, including Public Citizen, 
suggested the opposite. Public Citizen stated that NHTSA should require 
a longer period of post-crash recording once the trigger threshold is 
met. Public Citizen's rationale is that the NPRM's current time limit 
(0.5 second) would not record most data from rollover crashes (which 
typically last several seconds) or important post-crash information for 
non-rollover crashes. That organization also urged NHTSA to modify the 
``safety belt status'' data element to record from one second prior to 
an event to one second after an event. Public Citizen reasoned that 
this timing would allow crash investigators to determine whether the 
belt failed during a crash or whether the occupant intentionally 
unbuckled it after a crash.
    Advocates offered no specific technical comments for the EDR 
recording times and sampling rates. However, Advocates commented that 
the time durations must be sufficient to record the full event and 
provide adequate data, especially in rollover crashes. Advocates did 
acknowledge that there may be technological impediments or prohibitive 
costs to capturing data for the entire duration of full crash events. 
On the other hand,

[[Page 51020]]

IIHS indicated manufacturers may choose to forego or delay installing 
EDR features because of the significant costs that may be involved in 
recording extensive information on rollover angle, antilock brake 
activity, and stability control status. IIHS, for example, questioned 
the value of recording vehicle roll angle every 100 ms for one second 
before a crash.
    After carefully considering the comments, we have modified the 
recording intervals for a number of data elements. We made three basic 
modifications: (1) for the delta-V and acceleration data elements, we 
have changed the recording time from -0.1 to 500 ms, as proposed, to 0 
to 250 ms; (2) we changed the recording time from -8.0 to 0 sec, as 
proposed, to -5.0 to 0 sec for the following data elements: ``speed, 
vehicle indicated,'' ``engine throttle, % full,'' ``service brake, on/
off,'' ``engine, rpm,'' ``ABS activity,'' ``stability control,'' and 
``steering input'; (3) we changed the recording time for ``vehicle roll 
angle'' from -1.0 to 6 sec, as proposed, to -1.0 to 0.5 sec. Data 
sample rates in the final rule are unchanged from our proposal.
    Regarding the first modification, we changed the recording time for 
the delta-V and acceleration elements based on the comments. We agree 
with the commenters that recording these data elements for 500 ms 
challenges the microprocessing system, raising the risk of losing a 
complete crash record. We also believe that a lesser recording time 
would still be sufficient for our purposes. Further research conducted 
after our proposal indicates that the maximum delta-V will be reported 
95% percent of the time with a recording time of 250 ms.\39\ Our 
research also reveals that a 150 ms recording duration would not be 
sufficient.\40\ Based upon this information, we believe that a 250 ms 
recording time is sufficient for our purposes and also reduces the risk 
of losing EDR data because of a system malfunction.
---------------------------------------------------------------------------

    \39\ Niehoff, Peter, et al., Evaluation of Event Data Recorders 
in Full Systems Crash Tests, ESV Paper No. 05-0271 (2005).
    \40\ Id.
---------------------------------------------------------------------------

    We also reduced the recording time for several data elements from -
8.0 to 0 seconds to -5.0 to 0 seconds. We believe that this 
modification will further lessen the amount of data written to memory 
by the EDR and reduce the workload for the microprocessor. We do not 
believe that, for our purposes, the quality of data will significantly 
be reduced by changing the recording time.
    We have considered the comments concerning the recording interval 
for ``vehicle roll angle.'' In the NPRM, we proposed a recording 
interval from -1.0 to 6.0 seconds to allow for sufficient time to 
monitor the vehicle's roll angle after the crash event. We reasoned 
that recording up to 6 seconds after the crash event could be necessary 
because in cases where a frontal crash occurs, the vehicle might 
continue along some path for a second or more before it veers off the 
road and possibly overturns. We also considered SAE J1698-1, which 
classifies vehicle roll angle to be a high frequency data type with a 
recording interval from -300 ms up to 750 ms, and IEEE P1616, which 
specifies that ``roll rate'' and ``rollover'' data elements should be 
collected between -8 to 5 seconds.
    After carefully considering the comments concerning ``vehicle roll 
angle,'' we have modified the final rule by removing the specified 
recording interval for ``vehicle roll angle.'' We encourage vehicle 
manufacturers to use SAE J1698-1 and IEEE P1616 as a guideline for 
recording this data element. However, we have not included a specific 
recording interval in the final rule. We are providing flexibility to 
the automobile manufacturers that choose to record this data element. 
If we required a longer recording interval, it is possible that the 
costs would discourage automobile manufacturers from recording the data 
element. On the other hand, if we specified a shorter recording 
interval, we may not be provided with sufficient data for many crashes 
with subsequent overturns.
3. EDR Data Standardization (Format) Requirements
    The NPRM included a proposed section 563.8, which would require 
that the data elements listed in Tables I and II of the NPRM, be 
recorded in ``accordance with the range, accuracy, precision \41\, and 
filter class specified in Table III.'' GM, DaimlerChrysler, Ford, and 
Toyota submitted comments stating that these specifications (i.e., 
Table III of the NPRM) should not be included in the final rule. Other 
commenters suggested that the final rule should require data 
standardization and provided suggestions. Finally, GM, Ford, 
DaimlerChrysler, Toyota, Honda, and Delphi also made specific comments 
regarding the range, accuracy, and precision of acceleration data.
---------------------------------------------------------------------------

    \41\ This term was changed to ``resolution'' in the final rule. 
This change is technical, not substantive.
---------------------------------------------------------------------------

    GM believes that these parameters are beyond what is currently 
utilized in state-of-the-art EDRs to detect crashes, make deployment 
decisions, and record crash severity data, and GM argued that such 
provisions are not necessary to achieve the rule's safety benefits. 
DaimlerChrysler, Ford and Toyota provided different reasoning to reach 
the same conclusion as GM. They stated that NHTSA intended to use 
present design and performance capabilities of existing sensors, rather 
than to set new design and performance requirements. However, the 
current specification in paragraph 563.8 would run counter to that 
intent. Thus, DaimlerChrysler, Ford, and Toyota recommended that the 
range, accuracy, precision, and filter class be determined by the 
manufacturer for optimal restraint system performance, rather than EDR 
performance.
    Nissan, Mitsubishi, and ATA suggested that the final rule use the 
SAE J1698 resolution attribute instead of requiring specific levels of 
accuracy and precision. Nissan submitted a comment similar to other 
automakers, stating that the accuracy and precision standards do not 
correspond with current industry practice. If these accuracy and 
precision standards are retained, Nissan suggested that NHTSA should 
revise these standards to reflect present sensor performance 
specifications of each system feeding the EDR. In addition to Nissan, 
Mitsubishi requested that NHTSA consider SAE J1698 for the common 
output format for event data. Mitsubishi stated that many automobile 
manufacturers participated in creation of this standard to specify 
optimal standard output formats. Similarly, the ATA commented that it 
supports the use of standards developed by the SAE. As for filtering, 
Nissan questioned the rationale for requiring data to be filtered in 
accordance with SAE J211-1 before recording, instead of permitting 
filtering after data retrieval.
    Mr. Kast commented that, based on his studies of EDR data, some of 
the data elements are not recorded at the necessary resolution, 
accuracy, or duration to be of use (i.e., brake lights, acceleration, 
change of speed (computed delta-V), speed-vehicle indicated). He 
included a technical discussion of each element and the parameters 
necessary to acquire useful data.
    The ATA commented that data reliability must be assured. 
Specifically, ATA is concerned that inaccurate or erroneous data could 
result in incorrect assessments of the causes of accidents and of 
liability. ATA indicated that NHTSA should validate the technological 
ability to meet the

[[Page 51021]]

requirements as defined in Table III prior to any rulemaking. ATA 
questioned whether the data elements would be part of a certification 
process for a specified useful life or warranty period and whether 
service schedules would include the EDR.
    Several automobile manufacturers stated that the NPRM's required 
range and precision for accelerometers exceeds industry standards and 
are not currently commercially available. GM stated that the NPRM's 
requirements would have the effect of doubling the range and that 
increasing the accuracy would add significant costs not comprehended in 
the agency's cost estimates. GM currently utilizes 50 G 
accelerometers with an 8% accuracy. Ford, DaimlerChrysler, and Toyota 
also stated that accelerometers, as proposed, are not the industry 
norm, are not commercially available, and would increase the cost of 
compliance.
    Honda stated that the accuracy of mass-produced accelerometers used 
in motor vehicles can be near 10%. Honda requested that the 
final rule permit use of current G-sensors, which have ranges of 30 G 
to 50 G instead of 100 G. Honda argued that the NPRM had suggested 
these types of added costs were to be avoided. Honda also sought 
clarification as to whether '' * * * it is correct to say that the 
accuracy in Table III means only the recording error between the output 
value of each sensor to the recording unit and the input value to the 
retrieval tools?''
    Delphi made similar comments, stating that the NPRM's range and 
precision parameters for the longitudinal and lateral acceleration data 
elements are ``substantially different than [those] typically chosen 
for most crash sensing systems today.'' It stated that it would require 
manufacturers to change existing systems, potentially resulting in 
``sub-optimized system performance,'' to add separate sensors, 
resulting in increased costs, or require manufacturers to choose not to 
install an EDR. Delphi recommended that the normal acceleration element 
should not have fixed parameters for range, accuracy, precision, and 
sample rate. Instead, Delphi suggested that the value of those 
parameters should be reported as elements of the data record.
    Delphi commented that the accuracy requirements for accelerometers 
should allow a margin for sources of error attributable to other 
factors other than the accelerometers (e.g., alignment tolerances 
between the axes of the accelerometer and the vehicle). Delphi 
recommended that the accuracy requirement for longitudinal and lateral 
acceleration should be no less than +/-6 percent. Delphi recommended 
minimum limits of +/-50 G and 1 G be placed on the range and precision 
parameters for the longitudinal and lateral acceleration data elements, 
respectively and that the available range and precision of the sensors 
be reported as data elements.
    After carefully considering the comments, we have made a number of 
modifications to the range and accuracy requirements for the 
acceleration data elements. For these data elements, we proposed a 
range of ``100 G to +100 G and an accuracy of +/-1 G. In the final 
rule, the range and accuracy for the acceleration data elements is ``50 
G to +50 G with an accuracy of +/-5 percent. Based on our research \42\ 
and the comments, we believe that the new range and accuracy 
requirements are more realistic based upon what we now understand to be 
commercially available for vehicle production. Our research also leads 
us to believe that EDRs with accelerometers designed to meet these 
requirements will be sufficient to analyze safety equipment 
performance, a primary objective of this final rule.
---------------------------------------------------------------------------

    \42\ Niehoff, Peter, et al., Evaluation of Event Data Recorders 
in Full Systems Crash Tests, ESV Paper No. 05-0271 (2005).
---------------------------------------------------------------------------

    We have considered the recommendation off Mitsubishi and ATA that 
the final rule should use the SAE J1698 resolution attribute instead of 
requiring specific levels of accuracy and precision. After evaluating 
SAE J1698, we have concluded that the values in our proposal are nearly 
identical to or are less stringent than those found in SAE J1698. Thus, 
if an original equipment manufacturer were to use the SAE J1698 data 
resolution guidelines, they would be in compliance with the 
requirements of Table III.
4. EDR Data Retrieval and Whether To Require a Standardized Data 
Retrieval Tool/Universal Interface
    In the NPRM, we proposed requirements for EDR data retrieval (i.e., 
post-crash access to stored data). Under the NPRM's regulatory text 
(Sec. 563.12), the manufacturer of a motor vehicle equipped with an EDR 
would be required to furnish non-proprietary technical specifications 
at a level of detail sufficient to permit companies that manufacture 
diagnostic tools to develop and build a device capable of accessing, 
retrieving, interpreting, and converting the data stored in the EDR. 
The language would have required a manufacturer to submit the non-
proprietary technical specifications to NHTSA. We also requested 
comments on alternative approaches.
    Some commenters asked NHTSA to require standardization of data 
retrieval methods, arguing that a standardized data retrieval protocol 
will assist first responders and/or reduce cost. Other commenters 
suggested that we consider another approach, other than furnishing non-
proprietary technical specifications to NHTSA, to achieve the goal of 
making EDR retrieval tools available to crash investigators.
    SISC, ATA, SEMA, Advocates, and AAA recommended standardized 
retrieval methods so that emergency and first responder personnel can 
have quick and easy access to EDR data. SISC requested a standardized 
interface. SISC also believes that that retrieval of crash data in 
rural areas would be facilitated by the lower costs and easier access 
resulting from a single interface. For example, SISC suggested the U.S. 
Environmental Protection Agency (EPA) standards for onboard diagnostics 
of emission system performance. SEMA argued for a standardized 
retrieval method but indicated that the data should not be vehicle-
specific.
    AAA encouraged first responder access to data, but through ACN. AAA 
commented that transferal of accurate location coordinates, speed 
estimates, air bag deployment and other medically relevant information 
to EMTs should be encouraged through ACN. To that end, AAA commented 
that it supports NHTSA's proposal requiring manufacturers ``to provide 
sufficient technical detail to companies that manufacture commercial 
crash data retrieval systems.''
    Commenters offered other arguments for standardizing EDR data 
retrieval, including minimizing the ``tool-up'' costs and the 
inconvenience of having different types of data retrieval methods for 
each automaker. Three commenters referenced the On-Board Diagnostics 
(OBD) systems, requesting or opposing similar protocols for the EDR 
rule.
    PCIAA stated that the regulatory objective should be to avoid 
making EDR information access/retrieval more expensive and inconvenient 
than necessary. PCIAA commented that the failure to require 
standardization of the data retrieval method may preclude or diminish 
the opportunity for broader applications of the technology by the 
public and private sector. PCIAA further commented that NHTSA should 
adopt a standard in its final rule that minimizes new tool-up and 
licensing costs for the service and repair sectors. Because dealership 
service centers and independent automotive repair businesses have made 
significant

[[Page 51022]]

investments in recent years in scanner equipment to download or read 
data from the OBD electronic interface point, PCIAA urged NHTSA to 
consider requiring data retrieval through the OBD.
    NADA and SEMA made similar comments, asking NHTSA to consider 
setting standards for data retrieval communications protocols, 
connectors, and tools, similar to those of OBD systems. Additionally, 
SEMA argued that data access must include all data stored in the EDR, 
not just NHTSA-mandated data. That organization argued that the vehicle 
owner should be able to access all data stored in the EDR.
    On the other hand, Injury Sciences is opposed to relying on OBD 
protocols. Injury Sciences is opposed to relying on the connector as a 
means of retrieval because it asserts that data collection via the OBD 
works only if the electrical systems are intact. It argues that NHTSA 
should articulate retrieval requirements in the numerous instances when 
electrical systems are compromised and the extraction of data can only 
be accomplished from connecting directly to the device storing the 
information.
    Simplifying retrieval methods and minimizing costs were two common 
reasons suggested by commenters for standardizing EDR data retrieval, 
with some commenters providing technical suggestions for EDR data 
standardization. Public Citizen stated that NHTSA should standardize 
extraction protocols, technology, and interface location to ensure that 
data can be easily and quickly retrieved. Public Citizen believes that 
these steps would reduce overall costs. Advocates commented that the 
rule should require a uniform architecture for data retrieval. 
Advocates supported standardizing the retrieval method, citing higher 
costs for those retrieving data.
    Garthe Associates commented that the rule should require a uniform, 
non-contact retrieval method to rapidly and reliably download data. 
Garthe Associates suggested the use of radio frequency identification 
(RFID) or infrared (IR) for data retrieval. Garthe Associates also 
suggested specifications for the retrieval technology. Garthe 
Associates indicated numerous benefits of these technologies, including 
rapid access to crash data by EMS personnel. According to Garthe 
Associates, the estimated cost would be about $1/car.
    Mr. Fink stated that the rule should require standard software for 
downloading EDR data. He also commented that the same software and 
hardware should be able to access data from vehicle EDRs and commercial 
vehicle engine control modules. EPIC commented that the rule should 
address real-time data collection, which will become widely prevalent 
well before the proposed effective date for the rule.
    GM asked NHTSA to alter its proposal for data retrieval. In the 
NPRM, NHTSA proposed that each vehicle manufacturer must furnish non-
proprietary technical specifications at a level of detail sufficient to 
permit companies that manufacture diagnostic tools to develop and build 
a device capable of accessing, retrieving, interpreting, and converting 
the data stored in the EDR. GM recommended that we instead allow 
manufacturers to enter into a licensing agreement or provide other 
means for the tool(s) required for retrieving the EDR data. GM argued 
that aspects of EDR designs are often refined up to, and sometimes 
after, the start of vehicle production. GM argues that the provision 
would (1) potentially facilitate tampering with EDR data, (2) be 
impractical to accomplish at 90 days before the start of production, 
(3) result in a significant paperwork burden, (4) be unnecessary to 
satisfy a limited market for EDR download devices, and (5) require 
manufacturers to disclose proprietary information.
    Comments provided by DaimlerChrysler, Ford and Toyota were nearly 
identical to those of GM, except that they recommended that each 
manufacturer be required to certify to NHTSA that it has licensed the 
development of a download tool for each applicable vehicle. This is in 
contrast to GM, which was also open to other means of ensuring that a 
retrieval tool is available.
    SEMA commented that NHTSA should require manufacturers to provide 
information necessary for third parties to design and develop data 
access tools and should require the manufacturer to make the tools 
available to the public for a reasonable price and in a timely fashion.
    Ford stated that NHTSA should promulgate requirements that 
effectively prohibit tampering with EDRs and EDR data, because the 
value of EDRs is predicated upon the integrity of the data they 
contain.
    Honda commented that the NPRM's proposal to require the submission 
of data retrieval information no later than 90 days prior to the start 
of production of EDR-equipped vehicles is problematic. Honda argues 
that the modification or addition of information may become necessary 
near the start of production due to the detection of an inaccuracy or 
technical issue. Honda argued that under NHTSA's current proposal, a 
manufacturer would have to provide NHTSA with updated information and 
wait 90 days before it could start production with the modified EDR. 
Honda would like to be able to change the EDR specifications as soon as 
possible, and to produce vehicles equipped with the modified EDR as 
soon as possible. Therefore, Honda recommended that the final rule 
permit the submission of updated retrieval information as soon as it 
can be provided and for production of vehicles with the modified EDR to 
occur as soon as possible thereafter.
    ATA commented that specifications for the EDR interface should be 
provided to NHTSA but should not be part of the public domain. The ATA 
commented that a mandated, standard interface would threaten privacy 
rights. However, retrieval of data should be brief and should not 
impede the continued utilization, maintenance or repair of the subject 
vehicle.
    More than one commenter recommended changing the phrase in the 
``scope'' section of the regulatory text from ``it [the NPRM] also 
specifies requirements for vehicle manufacturers to make publicly 
available information that would enable crash investigators and 
researchers to retrieve data from EDRs' to ``it also specifies 
requirements for vehicle manufacturers to make commercially available 
tools and/or methods that enable crash investigators and researchers to 
retrieve data from EDRs.'' This change refers to the above comments 
that automobile manufacturers should be only required to make retrieval 
tools commercially available instead of having to furnish non-
proprietary technical specifications of the retrieval tools to the 
agency, as proposed in the NPRM.
    We have carefully considered the comments and recommended 
alternatives on this issue, and determined that an alternative approach 
will better meet the goal of ensuring that crash investigators are able 
to retrieve data from EDRs. We believe that requiring manufacturers to 
ensure by licensing agreement or other means that retrieval tools are 
commercially available will be sufficient for the purposes of this 
final rule. We believe that this revised approach will resolve concerns 
about the release of proprietary information. It will also result in 
less paperwork costs for the manufacturers.
    Under our revised approach, we are no longer following the NPRM's 
proposed requirement for vehicle manufacturers to submit information to 
the public docket to allow third parties to manufacture a retrieval 
tool for EDR data. Instead, the final rule requires

[[Page 51023]]

manufacturers and/or their licensees to make these tools commercially 
available. We expect that these retrieval tools will be accessible 
(i.e., for sale) for a reasonable period of time. That is, we 
anticipate that: (1) Retrieval tools will be available for several 
years after the vehicle whose EDR data it is designed to read has been 
sold, or (2) the capability to read EDR data for such vehicles will be 
integrated into a newer version of the tool, thereby making the new 
retrieval tool ``backward-compatible.'' (We note that current download 
tools designed for reading vehicle emissions-related data or engine-
control data have been designed to be backward-compatible, as has the 
Vetronix Crash Data Retrieval (CDR) tool for reading EDR data.) We 
anticipate that the movement toward backward-compatibility will 
continue and that there will be no issues associated with downloading 
EDR data from older vehicles covered by the EDR regulation set forth in 
this final rule. If this trend does not continue, the agency will 
consider appropriate action, as necessary.
    We are requiring the tool(s) to be commercially available not later 
than 90 days after the first sale of the motor vehicle for purposes 
other than resale. This addresses the timing concerns raised by 
commenters. Given that the retrieval tools will be commercially 
available, we do not believe it will be difficult to obtain information 
about how to obtain them.
    We have considered the comments asking us to require a standardized 
retrieval tool (or standardized retrieval software and hardware). In 
consideration of this issue, we assessed the comments concerning 
whether we should require a retrieval system similar to or utilizing 
the EPA/OBD protocols (68 FR 38427, June 27, 2003). However, such a 
requirement is beyond the scope of this rulemaking, since we did not 
propose to require a uniform retrieval tool in the NPRM. We do not 
believe that a uniform retrieval tool is necessary to achieve the 
purposes of this rulemaking. We believe that intended users will be 
able to access EDR data by our requiring manufacturers to ensure that 
retrieval tools are commercially available.
    We recognize, however, that there are potential benefits to 
standardizing the means of downloading EDR data. This could facilitate 
the future use of EDRs by first responders and possibly result in lower 
costs. This is an area that could potentially be addressed by voluntary 
organizations such as SAE and IEEE.
    We have considered NADA's and SEMA's comments that we should 
require access to all data stored in the EDR. However, we believe that 
it would not be appropriate to mandate the processing and storage for 
data that we currently have determined are not necessary for our goals 
of analyzing the performance of safety equipment, improving crash 
reconstruction, and enabling ACN.
    Additionally, we did not propose to require that vehicle owners 
have the ability of directly accessing EDR data. However, the 
requirement that vehicle manufacturers ensure that retrieval tools are 
commercially available should make it easier for vehicle owners to 
indirectly access stored EDR data.
    We considered the comments by Garthe and Siemens regarding a 
standardized, non-contact retrieval method. However, we did not propose 
the implementation of such technology in the NPRM, and will not include 
it in this final rule. Requiring automobile manufacturers to install a 
non-contact retrieval method is not necessary to achieve our stated 
purposes for this rulemaking.
    We have considered ATA's comments regarding access to EDR data, and 
we address this issue in our section on ``Privacy Issues.'' With regard 
to ATA's comments on mandating for brief retrieval we presently have 
not gathered sufficient information to mandate the brevity with which 
EDR data can be retrieved.
    We have considered the comments recommending that we address 
potential tampering with EDRs. We currently do not have information 
that leads us to believe that tampering with EDRs is a problem that 
necessitates us to develop requirements in this area. We may revisit 
this issue if we find that EDR tampering becomes a problem. However, we 
do believe one aspect of EDR design will discourage tampering. We are 
requiring that the captured file be locked for crashes that involve air 
bags. The locked file will be preserved and the file cannot be 
overwritten.
5. EDR Survivability and Crash Test Performance Requirements
    In the NPRM, we stated that if EDRs are to provide useful 
information, they must function properly during a crash, and that data 
must survive the crash. Accordingly, we proposed to require that EDRs 
meet specified requirements during and after the crash tests in FMVSS 
Nos. 208, 214, and 301. We also proposed that the data must be 
retrievable for not less than 30 days after the test and without 
external power. We chose not to propose more extensive survivability 
requirements, such as requiring EDRs to survive extreme crashes, fire, 
or fluid immersion.
    GM, DaimlerChrysler, Ford, Toyota, Nissan, and AIAM argued that in 
order to test for EDR survivability, as proposed in the NPRM, vehicles 
would have to be tested with engines running and various vehicle 
systems activated, presenting a danger to test personnel. Such tests 
also risk damaging test facilities, instrumentation, and photographic 
equipment resulting from fuel, oil, and/or battery fluid spillage.
    To solve this perceived problem, GM, DaimlerChrysler, Ford, and 
Toyota proposed an alternative approach to EDR crash survivability 
(i.e., a simulated laboratory test to verify EDR recording function and 
certification by engineering analysis to ensure sufficient energy 
reserve).\43\ According to the four commenters, the NPRM's current 
dynamic testing requirements for EDRs would greatly increase testing 
costs. GM, DaimlerChrysler, Ford, and Toyota also argued that the crash 
test provisions would not fulfill their intended purpose and that the 
provisions are unnecessary since the EDR function is typically co-
located in the restraint control module.
---------------------------------------------------------------------------

    \43\ Specifically, GM recommended replacing the proposed 
language in paragraph 563.10, Crash Test Performance and 
Survivability, with the following language:
    The data elements required by sec. 563.7 must be recorded so 
that they can be downloaded in the format specified by sec. 563.8; 
exist at the completion of the simulated test, and be retrievable by 
the methodology specified by the vehicle manufacturer under sec. 
563.12 for not less than 30 days after the simulated test, and the 
``complete data recorded'' element must read ``yes'' after the test. 
A simulated test for the purposes of this subsection consists of 
laboratory methods to provide data bus input representative of FMVSS 
[Nos.] 208 and 214 crash tests to the vehicle data bus, so that the 
EDR recording function can be verified. For those data elements not 
specified by FMVSS [Nos.] 208 and 214 (i.e., throttle angle, braking 
input, etc.), manufacturers will furnish simulated signals. In 
addition, manufacturers must certify through engineering analysis or 
other means that sufficient energy reserve exists in the subject 
module to ensure that all design-intended functions, including the 
deployment of restraint system components and the complete recording 
of EDR data elements as specified by this regulation, are fully 
supported in the event of power loss to the module from the 
vehicle's battery supply at any point following time zero, as 
defined by this regulation.
---------------------------------------------------------------------------

    Nissan stated that the NPRM's proposed regulatory text needs to be 
amended to reflect that engine RPM and throttle information will not be 
available in crash tests, which are performed without fuel. AIAM 
recommended clarifying the rule to indicate that EDR performance does 
not require the engine to be running and that, as a result, some data 
elements may not be recorded.

[[Page 51024]]

    GM, DaimlerChrysler, Ford, Toyota commented that storing crashed 
vehicles for 30 days following a test to ensure retrievability of data 
is impractical and unnecessary. These commenters stated that it is 
unreasonable to require data to be retrievable without external battery 
supply for 30 days, because current EDRs use external battery supplies 
to retrieve post-crash data. According to the commenters, the NPRM's 
requirements would necessitate adding a battery to the module, which 
would add significant cost and risk damage to the module circuitry due 
to electrolyte leakage. They also argued that this requirement for 30-
day retrievability is unnecessary to meet the safety purposes of 
furnishing additional data to aid in crash investigations.
    AIAM commented that the proposed regulatory text is not clear as to 
whether data must be retrievable without external power for up to 30 
days. AIAM suggested that the final rule should be clarified to require 
the EDR to store data without external power for up to 30 days but to 
permit an external power source for data retrieval.
    Nissan sought clarification for two issues related to 
survivability: (1) Whether an alternate power source would be required 
to ensure that the EDR is able to record up to 11 seconds of post-crash 
data; and (2) whether sensors would be expected to survive crashes to 
ensure delivery of data to the EDR. Mitsubishi stated that the final 
rule should not require data survivability in cases where there is a 
cut-off in the power supply or destruction of the electronic control 
unit. Mitsubishi argued that it is not technically feasible to require 
data recording if power is no longer directly supplied to the ECU.
    EPIC and the ATA made general comments regarding the survivability 
of EDRs. EPIC commented that EDR reliability is essential, ensuring 
that proper functioning of EDR systems becomes more critical as third 
parties (e.g., insurance companies and prosecutors) are provided access 
to EDR data. EPIC expressed concern that the level of survivability 
called for in the NPRM may not be sufficient to ensure reliable data. 
EPIC suggested text for the owner's manual encouraging owners to have 
the EDR inspected after a crash. ATA commented that EDRs must function 
properly during and after the specified crash tests.
    Several commenters gave specific suggestions for crash 
survivability. NADA commented that the rule should take into account 
EDR reparability and restoration. Advocates commented that the rule 
should require the EDR to be located in the passenger compartment in 
order to increase survivability. Hyundai and Kia commented that the 
rule should not require repositioning air bag control units to achieve 
crash survivability unless the repositioning would not adversely affect 
performance of the systems.
    Public Citizen and NTSB commented that the NPRM does not include 
requirements to ensure that the EDR will survive fire, fluid immersion, 
and severe crashes. To remedy this perceived deficiency, Public Citizen 
suggested that EDRs should be subjected to a rollover crash test or 
that they should meet survivability tests similar to those for airliner 
and locomotive ``black boxes.'' Public Citizen stated it is important 
that EDR data from severe crashes not be lost since such crashes may 
result in fatalities.
    We have carefully considered the comments regarding our testing 
requirements, and the commenters' position that requiring dynamic 
testing, as proposed, would be impracticable. After reviewing the 
comments from the manufacturers, we disagree that it is impracticable 
to require basic EDR crash survivability. However, we agree that 
certain proposed data elements cannot be recorded unless the crash 
tests are conducted with the engine running and vehicle systems 
activated. Those data elements are: ``Engine RPM'' and ``Engine 
Throttle % Full.'' At present, FMVSS Nos. 208 and 214 tests are not 
conducted with the engine running; compliance crash tests are only 
conducted with battery connected and vehicle systems activated. It was 
not our intention to propose any testing requirements beyond FMVSS Nos. 
208, 214, and 301. Testing with the engine running could create 
hazardous conditions for the test engineers. Therefore, we agree that 
``Engine RPM'' and ``Engine Throttle % Full'' cannot be recorded in 
current crash tests. We have modified the final rule to account for 
these concerns.
    As a result of our analysis of this issue, we have also realized 
that the braking input data element ``service brake on/off'' is not 
specified in FMVSS crash tests. Accordingly, there is no practical way 
to require manufacturers to test the survivability of this data element 
in the FMVSS Nos. 208 and 214 tests. Because there is no current way to 
test for these there elements, we have modified the final rule 
accordingly.
    After reviewing the comments, we believe that our proposal to 
require that data elements be retrievable for not less than 30 days 
after the test and without external power confused some commenters. We 
intended the proposed requirement that data be retrievable within 30 
days without external power to simply mean that the EDR data must be 
stored and saved in the system for at least 30 days without external 
power. This was not intended to mean that 30 days after the date of the 
crash, a crash investigator must be able to download the stored data 
with a download tool without an external power supply. We have modified 
the rule to clarify our original intentions.
    The final rule also modifies the number of days we will require EDR 
data to be retrievable after the crash test. Manufacturers have 
indicated that it usually takes three to seven days to complete the 
task of crash test data analysis and validation. Based upon this 
information, we believe that requiring that EDR data be retrievable up 
to 10 days better reflects the manufacturer's time frame of crash 
testing. We agree with manufacturers, based on this information, that a 
30-day requirement would require additional vehicle storage. 
Accordingly, we have modified the final rule.
    We have also considered the comments regarding EDR survivability in 
severe crashes or crashes involving fire or fluid immersion; however, 
we have not changed our position on requiring EDR survivability in 
these extreme cases. In the NPRM, we stated that EDR data from such 
crashes would be useful, but we do not have sufficient information to 
propose survivability requirements that would address such crashes. We 
also stated that countermeasures that would ensure the survivability of 
EDR data in fires might be costly. We have not engaged in research to 
promulgate survivability requirements for EDR data in these extreme 
cases. Moreover, we reiterate that the most important benefits of EDR 
data comes from enabling ACN and composite analysis, and we believe 
that this final rule will allow researchers to gather sufficient EDR 
data of statistical significance. We believe that we can meet the 
objectives of this rulemaking without requiring EDR survivability in 
extreme crashes.
    The comments of Ford, GM, Daimler Chrysler, and Toyota on EDR 
survivability also recommended deleting subsections (a)-(c) of the 
proposed regulatory text in Sec. 563.10. These commenters proposed an 
alternative testing protocol, as discussed above. The manufacturers 
recommended that a simulated test for data bus input of FMVSS Nos. 208 
and 214 be performed at room temperature and that the EDR data be 
stored at room temperature for 30 days after the tests.

[[Page 51025]]

We believe that testing requirements, as proposed by the manufacturers, 
would not be sufficient to meet our basic survivability requirements. 
These basic survivability requirements in the final rule, which will 
include the crash tests in FMVSS Nos. 208 and 214, are critical to 
verifying the performance and accuracy of EDRs because they reflect a 
controlled crash environment. The simulated tests recommended by the 
manufacturers for EDR crash survivability do not expose the EDR to a 
real crash environment. After carefully considering the comments, we 
believe that ensuring basic EDR survivability by requiring that EDRs 
meet specified requirements in accordance with FMVSS Nos. 208 and 214 
tests remains the best approach to ensure EDR survivability.
    We have, however, modified our crash test requirements in light of 
the comments we have received and in consideration of further 
information we have obtained. We have deleted the test associated with 
FMVSS No. 301. We believe that since most EDRs and other vehicle 
electrical systems are located in the front part of the vehicle, there 
is little chance that crash forces to the rear of the vehicle will 
affect EDR operation. Also, in the FMVSS No. 301 test, no air bags are 
deployed, so elements related to air bag deployment, that make up the 
vast majority the data collected by the EDR, are not collected.
    Also, we have decided not to require EDRs to meet requirements 
during crash tests listed under S13 of Sec.  571.208, as we proposed in 
the NPRM. The tests specified in S13 of Sec.  571.208 are currently 
subject to be gradually phased-out. After further consideration, we 
believe that the tests in FMVSS Nos. 208 and 214 will be sufficient to 
determine EDR survivability.
    The agency notes that in some FMVSS No. 214 tests (i.e., for 
vehicles equipped with only longitudinal delta-V sensors), the 
longitudinal trigger threshold may not be met because there may not be 
sufficient delta-V in that direction. For tests conducted pursuant to 
FMVSS No. 214, we would not expect the vehicle's EDR to record data 
unless the manufacturer records delta-V, lateral or any air bag 
(frontal, side, other) deploys.
    Our final rule represents tests that we believe will be sufficient 
to ensure basic EDR survivability. Furthermore, we would like to 
emphasize that this rule is not requiring any additional crash tests 
than what is currently required by existing FMVSSs. Tests for EDR crash 
survivability simply piggyback on test requirements for existing 
FMVSSs.
6. Compliance Date
    In the NPRM, we proposed an effective date of September 1, 2008 for 
the EDR regulation. We proposed this date with the intention of 
providing manufacturers adequate lead time to make design changes to 
their EDRs as part of their regular production cycle, minimizing costs. 
Almost all of the commenters on this issue believed that the proposed 
lead time was insufficient and/or would result in unnecessarily high 
costs, with most suggesting a phase-in beginning in 2008.
    GM, Ford, DaimlerChrysler, and Toyota stated if NHTSA issued a 
final rule for EDRs by September 1, 2005, that is consistent with their 
recommendations, they could support a four year phase-in beginning 
September 1, 2008 (10% of vehicle production at year 1, 25% at year 2, 
60% at year 3, and 100% at year 4). GM added that if the rule is 
appreciably different from its recommendations, it might need 
additional lead time to achieve compliance. GM reasoned that its 
recommended four-year phase-in would be an ``aggressive'' schedule 
because manufacturers would need to redesign, revalidate, and retool 
virtually every restraint control module, add greater power capability 
to those modules, and, in many cases, redesign the entire electrical 
architecture of the vehicle. Ford, DaimlerChrysler, and Toyota 
commented that their vehicle electrical/electronic architecture 
designs, which influence EDR feasibility, are presently being committed 
and cannot be readily changed for vehicles in model years before 2008. 
Hyundai and Kia commented that a four year phase-in period after the 
September 1, 2008 start date will be necessary to implement the design 
changes needed to meet the rule as proposed.
    Honda and Mitsubishi recommended that the effective date of the 
regulation should be no sooner than September 1st of the third year 
after publication of a final rule, with a phase-in period starting on 
that date. Honda's rationale is that it would be very difficult for all 
manufacturers to simultaneously develop and install compliant EDRs on 
all models by September 1, 2008.
    Subaru commented that the NPRM underestimates the time necessary 
for implementation. Because Subaru would have to acquire new memory 
devices, develop backup power sources, and possibly redesign its air 
bag system, Subaru requested additional lead time and a phase-in 
schedule for recording certain data elements. Subaru commented that its 
most state-of-the-art EDR technology is still not mature enough to meet 
all the proposed requirements. For example, its current air bag sensors 
do not meet the range and accuracy requirements. Subaru stated that it 
would probably remove all data recording rather than risk noncompliance 
if the rule were implemented as proposed.
    NADA commented that the rule should adopt a phased-in approach with 
multiple effective dates requiring that certain data recording 
capabilities be implemented in the near term, with additional data 
collection capabilities considered for the longer term. AIAM also 
commented that additional lead time would be necessary to meet the 
accuracy and precision requirements as proposed in the NPRM, due to the 
complexity of the required changes. AIAM suggested that the regulation 
should take effect with a pared down data set no sooner than the 
September 1st, three years after publication of the final rule and that 
the regulation should allow for a substantial phase-in period. If the 
final rule includes the complete set of proposed data elements, a 
longer lead time would be necessary. SISC commented we should provide 
sufficient lead time so that manufacturers can transition to multi-axis 
accelerometers (to ensure collection of lateral acceleration).
    We have considered the comments regarding our proposed effective 
date. Based upon the comments, we have decided to require covered 
vehicles manufactured on our after September 1, 2010 to comply with the 
requirements of this final rule. We believe that a lead time in excess 
of four years, particularly given the revised technical requirements, 
should prove adequate for all vehicle manufacturers and all vehicle 
lines, without the need for a phase-in. Vehicle manufacturers may 
voluntarily comply with the final rule prior to this date.
7. Privacy Issues
    The NPRM acknowledged that the recording of information by EDRs 
raises a number of potential privacy issues. These include the question 
of who owns the information that has been recorded, the circumstances 
under which other persons may obtain that information, and the purposes 
for which those other persons may use that information.
    In the NPRM, we stated that our rulemaking would not create any 
privacy problems. We explained that NHTSA would first obtain permission 
from the vehicle's owner before using the data. Furthermore, we believe 
that our objectives can be met by using a very brief snapshot of EDR 
data surrounding a crash. A broader use of

[[Page 51026]]

EDR data is not necessary for us to gather information or use EDR data.
    Many issues raised by commenters concerning privacy arise from the 
misconception that EDRs record data for prolonged intervals and 
personal information to study driver behavior. We noted in the NPRM 
that we were not proposing to require personal or location 
identification information. We also explained that we were proposing to 
standardize EDR data recording for an extremely short duration (i.e., a 
few seconds immediately before and after a crash). We did not propose 
to require data for prolonged recording intervals (i.e., several 
minutes) or audio/visual data that the public may associate with event 
data recorders in other modes of transportation, such as flight data 
recorders or locomotive event recorders. However, we note that another 
DOT agency, the Federal Motor Carrier Safety Administration (FMCSA), is 
currently engaged in rulemaking that would facilitate the use a 
different type of device, known as electronic on-board recorders 
(EOBRs), for documenting the hours of service of commercial drivers.
    In the NPRM, we expressed our sensitivity to privacy concerns, 
especially in relation to how we handle EDR data. We explained that 
NHTSA would first obtain a verbal release from the vehicle owner before 
using the data and fully comply with federal privacy law in its use of 
the information. Access to EDR data would not be affected by this 
rulemaking and would continue to be provided in limited situations. 
Furthermore, the design would most likely preclude public access to the 
EDR data because the interfaces will likely be located in the vehicle's 
passenger compartment.
    Some commenters argued that public safety outweighs any potential 
privacy issue or argued that privacy concerns were adequately addressed 
in the NPRM. Several individuals commented that the government and 
others will use EDRs to invade privacy. Still others identified privacy 
issues, but took differing positions on how to and who should address 
privacy concerns.
    GM, Ford, DaimlerChrysler, and Toyota commented that a FMVSS 
requiring EDR installation would permit the life-saving benefits of 
EDRs to be properly balanced, at the national level, with societal 
interests involving privacy and disclosure. These four commenters 
argued that unless there is Federal leadership, individual States will 
continue to regulate in the area of EDR privacy (e.g., through 
disclosure requirements). According to these automakers, unless this 
issue is dealt with comprehensively at the Federal level, the result 
could be a patchwork of State laws that would leave manufacturers in 
the untenable position of providing unique EDR systems and complying 
with disclosure language provisions on a State-by-State basis.
    EPIC commented that the NPRM inadequately protects the privacy of 
vehicle owners. According to EPIC, NHTSA has the responsibility to 
provide basic privacy protections and to clearly communicate to the 
public how EDR technology will be used. EPIC predicted that failure to 
do this would expose the rule to legal and political challenges. EPIC 
suggested that the rule should explicitly recognize the vehicle owner 
as the owner of EDR data. Moreover, EPIC expressed concern that many 
EDR systems currently record the complete VIN, including the serial 
number portion that can be used as a personal identifier.
    Several individuals commented on privacy and EDRs. Mr. Crutchfield, 
whose comments were representative of such commenters, expressed 
concern regarding the collection and use of EDR data. He argued that 
EDRs have no safety purpose and will be used to increase government 
revenues from fines, to increase rates or deny coverage by insurance 
companies, to justify seizure of private property, and to discriminate 
against individuals based on race, gender, age, regional origin, and 
socio-economic status.
    Mr. Leggett, an individual, commented on the collection and use of 
EDR data. He suggested that EDRs should be designed so that vehicle 
owners can remove them and that there should be no legal penalty for 
doing so. He also requested that the rule prohibit the use of EDR data 
in criminal and civil actions or by insurance companies. Mr. Leggett 
stated that the rule should specifically state that insurance companies 
may not require the use of EDRs, to ensure that the use of EDRs remains 
voluntary.
    Mr. King, an individual, commented that the rule should either 
provide protections for the vehicle owner (the presumptive data owner) 
or should be delayed until the passage of legislation addressing the 
issue. Mr. Lashway, along with fifty-two other individuals, commented 
that EDRs will be used to intrude into the privacy of individuals.
    Several commenters indicated that the ability to turn off or 
disable recording would resolve their concerns. Several also indicated 
that requiring written consent to acquire the data would be an 
acceptable solution. Some individuals commented that the EDR data are 
not reliable enough, thereby creating a danger to individuals 
confronted with countering the data in court. Commenters also suggested 
that vehicle purchasers should be provided with adequate notice about 
EDRs and EDR data at the time of first sale.
    SEMA commented that NHTSA should recognize that EDR data is the 
sole property of the vehicle owner. According to SEMA, a court order or 
consent of the vehicle owner should be required before EDR data may be 
released to insurance companies or before vehicle-specific data could 
be released to law enforcement. SEMA stated that an owner's consent 
could be provided prospectively via a form at the time of purchase 
(similar to current contracts for OnStar[supreg] subscriptions).
    Gelco commented that EDR data may contain personal information and 
may be easily accessible in the passenger compartment. Therefore, Gelco 
requested that the final rule explicitly or implicitly limit the access 
of the owner, lessor, or lessee to the data.
    The ATA commented that NHTSA should address privacy issues or 
coordinate with other appropriate Federal agencies to ensure that such 
issues are addressed. The ATA stated that it supports the practice of 
obtaining consent from the vehicle owner and commented that the data 
should be exempt from the Freedom of Information Act (FOIA). The ATA 
also expressed concern that a standard interface would make access to 
EDR data too easily accessible.
    Canada Safety Council commented that ownership of EDR data is 
unclear and that the issue needs to be resolved by legislators in the 
near future. The Council also commented that under the NPRM, emergency 
medical service personnel would not have easy access to crash severity 
data.
    Wyle Laboratories commented that NHTSA should consider 
certification of independent laboratories for EDR data management. 
Wyle's rationale is that such certification would facilitate data 
retrieval, validation, and distribution and would help protect the 
rights of each party with an interest in the data (e.g., manufacturer, 
owner, insurance carrier, regulator, and law enforcement agency).
    The ATA commented that, in contrast to what the NPRM states, much 
of the data is proprietary to the motor carrier or commercial vehicle 
operator. The ATA predicted that the volume of data that will be 
produced would: (1) Increase the likelihood that unskilled or untrained 
personnel would be involved in data analysis; (2) result in a 
misunderstanding or incorrect interpretation of data; (3) result in a 
use of erroneous data; and (4) lead to

[[Page 51027]]

obtaining and using data for purposes other than to improve vehicle, 
driver, and highway safety. Accordingly, the ATA suggested an 
appropriate level of training should be required to access, collect, 
and protect EDR data, especially considering the types and numbers of 
events that might warrant event data collection.
    AAA commented that law enforcement should have access to the data 
where a crash results in serious injury or fatality. AAA also commented 
that rules or laws need to be adopted to prohibit access to EDR data 
without a court order or permission from the owner. However, AAA did 
comment that EDR data that cannot be tied to a specific vehicle should 
be generally available for research purposes.
    National Motorists Association commented that it is inappropriate 
for EDR data to be used for criminal prosecutions and by insurance 
companies. The Association also expressed concern that EDR data is 
unreliable, which exacerbates the danger of its use for those purposes.
    Advocates commented that resolution of privacy issues should be 
left to the courts.
    Injury Sciences and Public Citizen did not view privacy concerns as 
an impediment to the EDR rule. Injury Sciences stated that it believes 
the NPRM provides adequate consideration and protection for the privacy 
of the individual. While acknowledging the importance of ensuring 
privacy, Public Citizen also did not see the EDR rule as raising a 
significant privacy concern. Public Citizen's comments suggested that 
``public health'' data provided by EDRs outweighs these privacy 
concerns. Public Citizen's rationale is that NHTSA already collects and 
uses EDR data, so the rule does not raise new privacy issues. 
Furthermore, Public Citizen stated that the NPRM addresses some 
existing privacy concerns by requiring a statement in the owner's 
manual to inform consumers as to the presence and role of the EDR in 
their vehicle.
    We have reviewed all of the comments regarding privacy and EDRs. As 
to comments concerning our planned use of EDR data, we hope that our 
continued efforts to educate and inform the public will help to correct 
any public misconceptions about the type of data that EDRs record and 
how that information is used.
    We stated in the NPRM that we are careful to protect privacy in our 
own use of EDR data. We obtain consent from the vehicle owner to gain 
access to EDR data. Furthermore, we assure the owner that all 
personally identifiable information will be held confidential. In 
handling this information, the agency does not make public any 
information contained in these records which has the potential to 
either directly or indirectly identify individuals, except as 
specifically required by law. Furthermore, prior to the release of 
information from databases containing EDR data (usually aggregated 
reports), the agency strips out the last six characters of the VIN 
(i.e., the portion that would allow identification of a specific 
vehicle and, potentially by indirect means, the identity of the 
vehicle's current owner). Therefore, we believe that the agency has 
taken adequate steps to ensure individual privacy vis-a-vis its use of 
EDR data.
    We understand that EDRs can generate concerns related to how EDR 
data are currently used or will be used by entities other than NHTSA. 
As we stated in the NPRM, our role in protecting privacy is a limited 
one. While we remain sensitive to the public debate about EDRs and the 
use of EDR data, we do not have statutory authority to address many 
privacy issues, which are generally matters of State and Federal law 
that we do not administer. These privacy issues were not created by 
this rulemaking (e.g., whether the vehicle owner owns the EDR data, how 
EDR data can be used/discovered in criminal/civil litigation, whether 
EDR data may be obtained by the police). EDRs have existed since the 
1970s, and our rulemaking on EDRs standardizes technology that has 
existed, in some cases, for decades.
    Other issues beyond the scope of this rulemaking include access to 
EDR data (including by law enforcement) and training of individuals to 
handle EDR data. As to Wylie Lab's comments, we did not propose 
certifying independent labs to handle downloaded EDR data for NHTSA, 
and we do not have a present need for such analysis.
    As noted earlier, we are not requiring or prohibiting on/off 
switches. Given that we are not requiring EDRs, we do not believe it 
would be appropriate to prohibit on/off switches. However, such 
switches could reduce the benefits from EDRs. Therefore, we believe it 
would be inappropriate to require such switches.
    We considered Mr. Leggett's comment concerning the reliability of 
EDRs in trials and other adjudicatory proceedings; however, we note 
that disputes about these issues are most appropriately resolved in 
individualized adjudications as needed.\44\ We are presently concerned 
with the reliability of EDR data only as it relates to our stated 
purposes of the analysis of safety equipment performance, 
reconstructing crashes, and fostering the development of ACN. We 
believe that the range, resolution, and accuracy standardization 
requirements are representative of current industry standards that are 
generally accepted in the industry, which we discussed in further 
detail above.
---------------------------------------------------------------------------

    \44\ See, e.g., Bachman v. General Motors Corp., 776 N.E.2d 262 
(Ill. App. Ct. 2002); Matos v. State, No. 4D03-2043 (Fla. Cir. Ct., 
Mar. 30, 2005); People v. Hopkins, No. 2004-0338 (N.Y. Co. Ct., Aug. 
30, 2004); Kevin Schlosser, ``Black Box'' Evidence, 231 N.Y. L. J. 
(Jan. 25, 2005).
---------------------------------------------------------------------------

    EDR technology continues to evolve, and public discussion about 
EDRs will continue. We hope to help address these concerns and foster 
continued acceptance of EDRs by requiring manufacturers of vehicles 
equipped with EDRs to include a standardized statement in the owner's 
manual, as discussed below. We also hope to establish an internet 
public education program to correct perceived public misunderstanding 
related to EDRs.
8. Owner's Manual Disclosure Statement
    In the NPRM, we proposed to require the following disclosure 
statement to be included in the owner's manual of vehicles that have an 
EDR:

    >This vehicle is equipped with an event data recorder. In the 
event of a crash, this device records data related to vehicle 
dynamics and safety systems for a short period of time, typically 30 
seconds or less. These data can help provide a better understanding 
of the circumstances in which crashes and injuries occur and lead to 
the designing of safer vehicles. This device does not collect or 
store personal information.

We proposed this disclosure statement in an effort to educate the 
public about EDRs, i.e., to inform consumers about the circumstances 
under which EDRs record data and the reasons why EDR data is collected.
    All commenters on this issue generally supported our proposal to 
require an EDR disclosure statement for consumers. We received several 
suggestions regarding the text and placement of that disclosure 
statement. Some thought that the language in the NPRM needed 
augmentation (or a complete rewrite) to address issues such as privacy, 
preemption, and ownership of and access to EDR data. We also received 
comments with proposed text to address telematic features, such as ACN, 
and specifically OnStar[reg].
    GM expressed support for requiring a standardized EDR disclosure 
statement in the owner's manual. However, GM recommended expanding the 
statement to more fully inform consumers (e.g., by

[[Page 51028]]

providing examples of the type of information recorded, explanation 
that no recording occurs under normal driving conditions, and an 
explanation of download protocols) and to respond to issues currently 
being addressed at the State level (e.g., access to EDR data). In light 
of the above, GM also suggested that the disclosure statement should 
inform consumers if their vehicle is equipped with a telematic system 
that may collect personal and/or vehicle information. GM recommended 
the following disclosure statement:

    This vehicle is equipped with an event data recorder (EDR). The 
main purpose of an EDR is to record, in certain crash or near crash-
like situations, such as an air bag deployment or hard braking, data 
that will assist in understanding how a vehicle's systems performed. 
The EDR is designed to record data related to vehicle dynamics and 
safety systems for a short period of time, typically 30 seconds or 
less. The EDR in this vehicle is designed to record such data as:
     How various systems in your vehicle were operating;
     Whether or not the driver and passenger safety belts 
where buckled/fastened;
     How far (if at all) the driver was depressing the 
accelerator and/or brake pedal; and,
     How fast the vehicle was traveling.
    These data can help provide a better understanding of the 
circumstances in which crashes and injuries occur. NOTE: EDR data 
are recorded by your vehicle only if a crash or near crash situation 
occurs; no data are recorded by the EDR under normal driving 
conditions.
    To read data recorded by an EDR, special equipment is required 
and access to the vehicle or the EDR is required. In addition to the 
vehicle manufacturer, other parties, such as law enforcement, that 
have the special equipment, can read the information if they have 
access to the vehicle or the EDR.
    [If the vehicle is equipped with telematic system(s), the 
following statement must also be included in the owner's manual.]
    Your vehicle may be equipped with onboard telematics that 
provide safety and convenience services such as GPS-based navigation 
or cellular wireless connectivity, and your vehicle may collect 
personal or vehicle information to provide such services. Please 
check the service's subscription agreement or manual for information 
about its data collection.

    According to GM, the NPRM's owner's manual language may not be 
sufficient to obviate or to preempt current or future State disclosure 
requirements. GM's recommended disclosure statement also omits 
reference to ``personal information,'' as we proposed in the NPRM, 
because GM believes that phrase is potentially ambiguous.
    Comments from DaimlerChrysler, Ford, and Toyota were similar to 
GM's comments, although they differed in two areas. Each recommended 
that the EDR rule permit vehicle manufacturers to supplement the 
required language with additional information that the manufacturers 
deem appropriate for their respective vehicle designs. Each also 
omitted the language GM included related to telematic systems.
    SEMA, Advocates, and Mr. Bruce Funderberg commented that customers 
should be notified if a vehicle is equipped with an EDR prior to 
purchasing the vehicle. SEMA stated that vehicle dealers should be 
required to notify consumers about EDRs, consistent with State and 
local laws and that subscription services (e.g., OnStar[supreg]) should 
be required to notify purchasers of the types of EDR information that 
may be transmitted and to whom the data would be provided.\45\ 
According to Advocates, NHTSA should require dealers to provide a copy 
of the statement to purchasers at the time of sale along with a 
brochure written in both English and Spanish. Advocates also supported 
the use of additional methods to educate the public about EDRs, such as 
public service announcements, agency publications, and NHTSA's Web 
site. Mr. Funderburg, an individual, commented on vehicle owners' lack 
of knowledge about EDRs, suggesting that manufacturers need to provide 
better notice to purchasers about EDRs. He also recommended that the 
EDR should be optional equipment that purchasers may decline.
---------------------------------------------------------------------------

    \45\ SEMA suggested the following disclosure language:
    This recorded data may not be retrieved or downloaded by anyone 
other than the owner of the vehicle except in certain specific 
circumstances: (1) With the consent of the owner; (2) by court 
order; (3) by an authorized person for purposes related to improving 
vehicle safety provided the identity of the registered owner or 
driver is not disclosed and the information is of a non-vehicle 
specific nature; or (4) the data is retrieved for the purpose of 
determining the need or facilitation of emergency medical response.
    In cases where vehicles are equipped with a recording device as 
part of a subscription service, the fact that information may be 
recorded or transmitted must be disclosed in the subscription 
service agreement.
---------------------------------------------------------------------------

    EPIC commented that the notice to owners should be more specific 
about the ownership of and access to EDR data. EPIC also commented that 
ACN systems present unique privacy issues, stating ``for EDRs that use 
communications systems--such as OnStar[supreg], which uses wireless 
phone networks--the EDR should not initiate communication unless an 
accident is detected or if the driver uses a manual feature to initiate 
communications for purposes of transmitting driving data.'' \46\ EPIC 
commented:
---------------------------------------------------------------------------

    \46\ EPIC proposed the following additional text for the 
statement in the owner's manual for vehicles that contain ACN or an 
EDR connected to a communications network:
    The event data recorder is connected to a communication system 
capable of automatically contacting emergency services when it 
detects an accident. The event data recorder will only initiate 
communication in the event of an accident or if the driver uses the 
manual feature to initiate communication with either emergency 
services or the communications provider (e.g., for a service that 
provides driving directions from an operator).

    Consent of the vehicle owner should be required for the 
disclosure of EDR driving data to the NHTSA or any other government 
or commercial organization, including automotive insurance 
companies. Such consent should be fully consensual, meaning for 
example that automotive insurance contracts should not be 
conditioned upon access to EDR data.\47\
---------------------------------------------------------------------------

    \47\ EPIC proposed the following additional text for the 
statement in the owner's manual for vehicles that contain ACN or EDR 
connected to a communications network:
    Your consent is required for the data to be disclosed to the 
National Highway Traffic Safety Administration--a federal agency 
that gathers information about traffic accidents to improve vehicle 
and road safety--or any other government or private organization, 
including automotive insurance companies.
    EPIC also commented that if a partial VIN is included in EDR, 
the following text should be added to the owner's manual:
    Only the part of your vehicle identification number (VIN) that 
includes information about the make and model of your vehicle will 
be collected by the event data recorder. The unique serial number 
portion of the VIN will not be collected.

    In addition, EPIC commented that the vehicle owner should be 
instructed to have the EDR inspected if the vehicle has been involved 
in an accident, flooding, or fire.\48\
---------------------------------------------------------------------------

    \48\ EPIC proposed the following additional text for the 
statement in the owner's manual:
    If your vehicle has been involved in a serious accident or has 
been subject to flooding or fire, your event data recorder may have 
been damaged. If it was involved in one of these situations, please 
have your event data recorder inspected by an authorized dealer.
---------------------------------------------------------------------------

    The National Motorists Association, Advocates, AAA, and ATA all 
made comments that the proposed disclosure statement is inadequate to 
address an array of consumer concerns, and some suggested alternative 
language. PCIAA commented that the required, specific disclosure 
statement proposed in the NPRM is inadequate because the statement 
could become obsolete quickly and because vehicle owners rarely refer 
to or use their owner's manual. Advocates commented that the required 
statement in the owner's manual is necessary but not sufficient to 
educate the public about EDRs and address privacy concerns. AAA 
commented that there is insufficient consumer notification about access 
to

[[Page 51029]]

EDR data, stating that manufacturers should disclose in the owner's 
manual whether any outside parties that have access to the data and 
under what circumstances the data are shared. ATA commented that the 
statement in the owner's manual should disclose that an EDR is present 
and that the EDR does not collect or store personal information. The 
ATA also stated that additional public information would be desirable.
    After considering the public comments, we have decided to adopt a 
more detailed disclosure statement, along the lines recommended by GM, 
DaimlerChrysler, Ford, and Toyota. We believe that the more detailed 
statement will provide consumers with a fuller understanding of the EDR 
installed in their vehicles.
    However, we are not adopting the recommended language in GM's 
comments related to telematic systems, because such systems are not 
directly the subject of this rulemaking. We note that the comments of 
DaimlerChrysler, Ford, and Toyota did not include language related to 
telematic systems, although the balance of their recommended disclosure 
statements were virtually identical to that of GM. The capabilities of 
telematic systems and the level of integration between such systems and 
the EDR may also vary depending upon the given technology. For these 
reasons, we have decided not to require language in the specified 
disclosure statement on telematic systems. However, vehicle 
manufacturers may include a discussion of applicable telematic systems 
in the vehicle owner's manual, if they choose to do so.
    In addition, we note that we are permitting vehicle manufacturers 
to supplement the required owner's manual statement on EDRs with 
additional information, if they choose to do so. Vehicle manufacturers 
will have specific knowledge about their EDRs, and in some situations, 
vehicle owners may benefit from such additional information.
    In response to SEMA's comment that vehicle dealers should also be 
required to notify consumers about EDRs and Advocates comment 
requesting an additional brochure, we believe that such requirements 
would be largely redundant of the information required in the owner's 
manual, and hence unnecessary.
    In addition, we have decided not to adopt SEMA's recommendation for 
a requirement for subscription services, such as OnStar[supreg], to 
disclose information about the types of data that may be transmitted 
and to whom they may be transmitted, for the following reasons. First, 
the regulation of such services is outside the scope of this 
rulemaking, and second, consumers are generally made aware of such 
services up-front, particularly where they must pay a fee for the 
continuation of service. To the extent that consumers are concerned 
about the data gathered or reported by these services, they are free to 
pose such questions to the provider.
    Regarding Mr. Funderburg's comments that EDRs should be optional 
equipment that purchasers may decline, we note that making EDRs an 
option could add unnecessary production costs. Moreover, there are no 
benefits associated with not having an EDR. Furthermore, taking such a 
position would run counter to our safety goals of securing more and 
better EDR data and enabling ACN.
    For the reasons discussed more fully under section IV.B.7 of this 
document, we do not believe that EDRs raise meritorious privacy 
concerns, because they do not collect individual identifier 
information. We believe that the disclosure statement we have adopted 
provides a clear picture of the types of data collected by EDRs and the 
intended uses of that data.
    We have decided not to adopt EPIC's recommended language warning 
the consumer to have the EDR inspected after the vehicle is in a crash 
or is subject to fire or flooding. We do not believe that such language 
is necessary, because in such cases, the vehicle owner will normally 
have the vehicle examined by both an insurance adjuster and an 
automotive repair expert, professionals who will diagnose resulting 
problems with all vehicle systems, including the EDR.
    In response to commenters who argued that our proposed owner's 
manual disclosure statement is inadequate because it is too limited, we 
note that under the final rule, we are requiring an expanded disclosure 
statement. We believe that our specified owner's manual disclosure 
statement provides adequate notice as to the presence and function of 
the EDR.
    We have considered the comments arguing that our proposed owner's 
manual statement could become quickly obsolete. NHTSA intends to 
closely follow the development of EDR technology. If we determine that 
these devices have evolved in such a way as to render our disclosure 
statement inadequate, we would consider how to amend the required 
language. In addition, as stated above, we are permitting vehicle 
manufacturers to augment the required disclosure statement with 
additional information based upon the specifics of the EDRs installed 
on the vehicle. For these reasons, we believe that the EDR-related 
information provided to consumers will be sufficient for most 
consumers.
9. Preemption
    GM, DaimlerChrysler, Ford, Toyota, AIAM, and NADA recommend that 
the final rule for EDRs should explicitly state that it preempts 
inconsistent State and local regulations. GM is concerned that without 
a clear statement of the preemptive effect of the final rule, 
manufacturers could be faced with a patchwork of State and local 
requirements. AIAM expressed concern that the failure to preempt 
inconsistent State and local regulations could result in manufacturers 
being required to provide limited, circumscribed, or deactivated EDR 
systems and inconsistent disclosure/owner's manual language on a State-
by-State basis. AIAM argued that the consistency across the nation 
would aid in the public acceptance of EDRs and would help keep costs 
down. NADA commented that the rule should expressly reference the 
degree to which inconsistent State or local regulations are preempted.
    We have considered the comments concerning the preemption of 
conflicting State regulations and agree that a patchwork of State laws 
is not desirable. We expect that general principles of preemption law 
would operate so as to displace any conflicting State law or 
regulations.
    It is our view that any State laws or regulations that would 
require or prohibit the types of EDRs addressed by our regulation, or 
that would affect their design or operation, would create a conflict 
and therefore be preempted. Specifically, this would include State EDR 
technical requirements, such as ones requiring EDRs in motor vehicles 
(except for State-owned vehicles), requiring that EDRs record specific 
data elements, and/or requiring EDRs to meet specific technical 
performance or survivability requirements.
    Further, it is our view that any State laws or regulations that 
imposed, for the types of EDRs addressed by our regulation, additional 
disclosure requirements on vehicle manufacturers or dealers would 
likewise create a conflict and therefore be preempted. We have devised 
an appropriate statement for the owner's manual to make the operator 
aware of the presence, function, and capabilities of the EDR. 
Inconsistent or additional State disclosure requirements would 
frustrate the purposes of our regulation by potentially creating 
confusion or

[[Page 51030]]

information overload, thereby reducing the benefit of the required 
statement. The need to meet different disclosure requirements for 
different States would also increase costs, making it less likely that 
manufacturers would provide EDRs.
    It is our intent to provide one consistent set of requirements, 
including a specified statement in the owner's manual, for vehicle 
manufacturers that choose to install EDRs. We believe that this 
approach will enhance the quality of EDR data by standardizing the 
content, format, and accuracy of such data, thereby increasing its 
comparability and overall usefulness; we further believe that the 
standardized data will be of greater benefit for safety equipment 
analysis and crash reconstruction. We also believe that this minimum 
data set provides key elements in a standardized format that will 
foster the development of ACN and other telematic systems.
    We believe that State laws inconsistent with this final rule would 
frustrate the final rule's purposes. For example, additional State 
requirements would increase the costs of EDRs and make it less likely 
that manufacturers would voluntarily provide them. Additional State 
requirements could also hamper the development of future EDRs by 
pushing their development in ways that are not optimal for safety. 
Among other things, given limitations in data processing capabilities, 
requirements for additional data elements could make EDRs less 
effective in real world crashes in recording the data elements NHTSA 
has determined to be most important. (As discussed in section IV.B.2 of 
this notice, we believe that recording of additional data elements, 
which are currently of lesser value for our stated purposes, would not 
only result in significantly higher costs but would also risk 
overburdening the microprocessing and memory capabilities of EDRs. This 
could increase data recording times, and it could also increase the 
risk of system failure, potentially resulting in the loss of all EDR 
data.)
    In addressing the issue of preemption, we note that the effective 
date for our EDR regulation is 60 days after publication of this rule, 
and that the compliance date is September 1, 2010. It is our view that 
our regulation has preemptive effect between the effective date and 
September 1, 2010, as well as after that latter date. In New Jersey 
State Chamber of Commerce v. State of New Jersey,\49\ the Court held 
that a delay in the start-up date of certain provisions of the 
Occupational Safety and Health Administration's (OSHA's) Revised 
Asbestos Standards did not affect the effective date of preemption, in 
that case upon publication in the Federal Register (holding that 
preemption arises before the regulation becomes operative, in cases 
where an agency provides additional time for regulated entities to take 
steps to prepare for compliance). The same principle applies here, and 
we have a substantive reason for structuring the effective date and 
compliance date in the manner we have done. Once the EDR regulation is 
effective, a conflict with an inconsistent State law would arise 
immediately and impact achievement of our ultimate objectives for 
compliance in 2010.
---------------------------------------------------------------------------

    \49\ 653 F.Supp. 1453, 1462 (D. N.J. 1987).
---------------------------------------------------------------------------

    Specifically, we selected this compliance date to provide 
sufficient lead time to enable manufacturers to incorporate necessary 
changes as part of their routine production schedules. Thus, we expect 
that, in order to meet the requirements of our regulation, between now 
and September 1, 2010, vehicle manufacturers will be gradually 
redesigning their EDRs, modifying vehicle systems and components that 
feed into EDRs, and integrating EDRs into numerous models of vehicles. 
Furthermore, a vehicle manufacturer may begin complying with the EDR 
regulation once it becomes effective. Thus, any State laws or 
regulations that would require or prohibit the types of EDRs addressed 
by our regulation, or that would affect their design, or that would 
establish a compliance date earlier than September 1, 2010, would 
conflict with and frustrate the purposes of our regulation. Among other 
things, such laws or regulations would interfere with the process of 
manufacturers gradually redesigning their EDRs, modifying related 
vehicle systems and components, and integrating EDRs into vehicles in 
order to meet our requirements during that timeframe.
    The agency is aware of ten States that have passed laws relating to 
EDRs in the fields preempted by this final rule.\50\ Most of these 
States require that the vehicle purchaser be notified that the motor 
vehicle is equipped with an EDR. Three States, Arkansas, Colorado, and 
North Dakota, require additional information. Of those three States, 
Arkansas and North Dakota have the broadest disclosure requirements. 
Arkansas requires disclosure of the presence of the EDR, the type of 
EDR, and the type of data that is recorded, stored, or transmitted.\51\ 
North Dakota requires disclosure of the presence, capacity, and 
capabilities of the EDR.\52\
---------------------------------------------------------------------------

    \50\ Arkansas, California, Colorado, Maine, New Hampshire, New 
York, Nevada, North Dakota, Texas, and Virginia.
    \51\ Arkansas Code, Title 27, Chapter 37, Subchapter 1, Section 
103.
    \52\ North Dakota Century Code, Title 51 Sales and Exchanges, 
51-07-28.
---------------------------------------------------------------------------

    We believe that the statements meeting our disclosure requirement 
in the final rule would satisfy even the broadest of the existing State 
disclosure requirements. Further, it does not appear that any of the 
existing State requirements regarding disclosure would conflict with 
the final rule.
    This rule does not address certain other issues generally within 
the realm of State law, such as whether the vehicle owner owns the EDR 
data, how EDR data can be used/discovered in civil litigation, how EDR 
data may be used in criminal proceedings, whether EDR data may be 
obtained by the police without a warrant, whether EDR data may be 
developed into a driver-monitoring tool, and the nature and extent that 
private parties (including insurance companies, car rental companies, 
and automobile manufacturers) will have or may contract for access to 
EDR data. These issues are instead being addressed by State 
legislatures.
10. Applicability of the EDR Rule to Multi-Stage Vehicles
    In the NPRM, we stated that our proposed EDR rule would apply to 
the same vehicles that are required by statute and by FMVSS No. 208 to 
be equipped with frontal air bags (i.e., passenger cars, multipurpose 
passenger vehicles, trucks, and buses with a GVWR of 3,855 kg or less 
and an unloaded vehicle weight of 2,495 kg or less, except for walk-in 
van-type trucks or vehicles designed to be sold exclusively to the U.S. 
Postal Service). This covers most light vehicles, including multi-stage 
vehicles. We believe applying this rule to all vehicles that are 
currently subject to FMVSS No. 208 is appropriate since most EDRs are 
closely associated with frontal air bags and all of these vehicles must 
meet the advanced air bag requirements of FMVSS No. 208, which will be 
completely phased in by manufacturers before compliance with this final 
rule is required.
    Several commenters suggested changing our proposal to provide an 
exception for multi-stage vehicles and incomplete, intermediate, and 
final stage manufacturers. GM, DaimlerChrysler, Ford, and Toyota 
expressed support for either excluding incomplete, intermediate, and 
final

[[Page 51031]]

stage manufacturers from the requirements of the rule by specifically 
excluding these manufacturers in the regulatory text or by requiring 
those manufacturers to certify compliance one year after the last 
applicable date for manufacturer certification of compliance provided 
under the final rule. GM's point is that the proposed EDR rule would 
result in a significant burden on incomplete, intermediate, and final 
stage manufacturers. GM argues that the integration of EDR functions 
into a vehicle is a complex task requiring advanced communications and 
data processing technologies that may be beyond the capabilities of 
many small businesses.
    ATA asserted NHTSA has not involved final stage vehicle 
manufacturers or accessory installers in an appropriate dialog. ATA 
encouraged NHTSA to conclude that there is no possibility that EDR 
performance could be affected during any type of completion or 
conversion or accessory installation. On the issue of the effect of the 
EDR requirements on altered vehicles, NADA commented that NHTSA should 
``consider the complexities that may be involved for light-duty 
vehicles manufactured in two or more stages or which are altered prior 
to first sale.''
    We have considered the comments that we provide an exception or 
otherwise delay the effective date of this rulemaking for incomplete, 
intermediate, and final stage manufacturers (i.e., multi-stage 
vehicles). Since the NPRM was published, NHTSA has issued a final rule 
pertaining to certification requirements for vehicles built in two or 
more stages and altered vehicles (see 70 FR 7414 (February 14, 2005)). 
The amendments made in that final rule become effective September 1, 
2006. In relevant part, the multi-stage certification final rule 
amended 49 CFR 571.8, Effective Date, and it added a new subparagraph 
(b) providing as follows:

    (b) Vehicles built in two or more stages vehicles and altered 
vehicles. Unless Congress directs or the agency expressly determines 
that this paragraph does not apply, the date for manufacturer 
certification of compliance with any standard, or amendment to a 
standard, that is issued on or after September 1, 2006 is, insofar 
as its application to intermediate and final-stage manufacturers and 
alterers is concerned, one year after the last applicable date for 
manufacturer certification of compliance. Nothing in this provision 
shall be construed as prohibiting earlier compliance with the 
standard or amendment or as precluding NHTSA from extending a 
compliance effective date for intermediate and final-stage 
manufacturers and alterers by more than one year.

    In light of the agency's policy on multi-stage manufacturer 
certification, as expressed in the February 14, 2005 final rule, we 
have decided to apply that principle to the compliance date for final-
stage manufacturers and alterers. Thus, final-stage manufacturers and 
alterers must comply with this rule for vehicles manufactured on or 
after September 1, 2011. However, final-stage manufacturers and 
alterers may voluntarily certify compliance with the standard prior to 
this date.
11. Applicability of the EDR Rule to Heavy Vehicles and Buses
    In addition to multi-stage vehicles, Public Citizen and Advocates 
commented that NHTSA should extend the rule's applicability to include 
other vehicles, such as heavier trucks and 15-passenger vans. Public 
Citizen commented that all new vehicles, including large trucks, should 
be required to be equipped with EDRs, and the organization encouraged 
NHTSA to undertake a separate rulemaking to require EDRs in large 
trucks. Public Citizen stated that the benefit realized by EDRs is 
directly proportional to the number of vehicles equipped with these 
devices and that full fleet penetration is critical to the accuracy and 
utility of EDR data. Public Citizen further commented that an EDR 
requirement for large trucks could help improve industry practices and 
driver behavior. Similarly, Advocates commented that the rule should 
include 15 passenger vans and heavier light trucks because those 
vehicles have relatively high rollover rates, high risk of injury to 
multiple occupants, and are exempt from other safety regulations (e.g., 
side impact and roof crush resistance).
    While EDR requirements for heavier vehicles are outside the scope 
of this rulemaking, we note that many 15-passenger vans are within the 
applicable weight range for this final rule, and thus, are required to 
comply with the EDR regulation. Further, we note that some original 
equipment manufacturers, such as GM, are installing EDRs in their 
medium trucks equipped with air bags.
    As noted in the NPRM, we are not addressing in this document what 
future role the agency may take related to the continued development 
and installation of EDRs in heavy vehicles. We will consider that topic 
separately, after consultation with the Federal Motor Carrier Safety 
Administration. As noted previously, FMCSA is currently engaged in 
rulemaking that would facilitate the use of Electronic On-Board 
Recorders for recording and documenting the hours of service of 
commercial drivers.
    We believe that deferring consideration of requirements for EDRs 
installed on heavy trucks is appropriate for the following reasons.
    First, it would provide the agency with time to build experience in 
terms of standardization of EDR data in light vehicles. This experience 
could then be applied to our consideration of heavy trucks.
    Second, because the relevant data to be gathered by EDRs installed 
in heavy trucks are not identical to that of light vehicles, we believe 
any such requirements should come in a separate regulation.
    Third, because EDRs in light vehicles rely heavily upon sensors and 
diagnostic equipment associated with the vehicle's air bag system, the 
agency must carefully assess the costs, benefits, and lead time 
necessary for EDR requirements for heavy trucks, which may not have 
systems with all the necessary hardware. We understand that heavy truck 
manufacturers, suppliers, and others are engaged in EDR-related efforts 
with SAE, which will result in recommended practices for these devices. 
NHTSA is closely monitoring these efforts by the SAE working group. 
NHTSA is also closely following activities in other governmental 
agencies, including FMCSA and NTSB.
    Finally, separate consideration of EDR requirements for heavy 
trucks will expedite promulgation of this final rule for EDRs in light 
vehicles, thereby encouraging further positive developments based upon 
standardized EDR data.
12. Automatic Crash Notification and E-911
    The NPRM stated that the purpose of this rulemaking is to help 
ensure that EDRs record, in a readily usable manner, data necessary for 
effective crash investigations, analysis of safety equipment 
performance, and automatic crash notification systems. It is NHTSA's 
position that this data will help provide a better understanding of the 
circumstances in which crashes and injuries occur and will lead to the 
designing of safer vehicles.
    Including ACN as a stated purpose of the EDR rule drew comments. 
Commenters recommended additional clarifying language or deleting 
relevant portions of the proposed regulatory text so that the rule 
cannot be construed as a limitation on manufacturers' ability to offer 
telematics features, such as ACN. GM, Ford, and Toyota recommended that 
the final rule expressly state that it does not limit manufacturers' 
ability to

[[Page 51032]]

offer ACN and other telematics features. Likewise, PCIAA commented the 
rule should not ``preclude EDRs and similar vehicle technology (i.e., 
intelligent vehicle systems-telematics) from being fully leveraged by 
the public and private sectors.''
    GM argued that because ACN is not being proposed in this 
rulemaking, the language referencing ACN should be dropped from the 
regulatory text. GM further argued that the proposed EDR rule makes no 
provision for the software, hardware, and infrastructure required to 
make use of ACN-related data. DaimlerChrysler made a similar comment, 
adding that ACN infrastructure was last estimated to cover only 25% of 
the United States, principally in urban areas. DaimlerChrysler stated 
that benefits of ACN, other than those related to better crash data, 
are speculative and out-of-scope.
    We acknowledge that this final rule does not regulate or require 
ACN systems. Nonetheless, we are retaining ACN as a stated reason to 
require EDR data standardization because we believe that the final rule 
would have ancillary benefits, such as facilitating ACN development. 
However, our other stated purposes fully justify the rule. We emphasize 
that this final rule does not limit the ability of manufacturers to 
offer ACN or other telematics devices.
    To reiterate our earlier reasoning, we note that the NPRM provides 
a detailed explanation of the relationship between EDRs and ACN 
systems. In addition, the ENHANCE 9-1-1 Act of 2004 requires the 
Department of Transportation to help coordinate and to speed the 
deployment of Wireless Enhanced 9-1-1. ACN has the potential for 
interfacing with nation-wide Wireless Enhanced 9-1-1 deployment by 
providing immediate and accurate crash location information to Public 
Safety Answering Points. This will expedite the dispatch of emergency 
services to the crash scene, help ensure that EMS personnel can locate 
the crash, and speed the provision of lifesaving emergency medical 
services to traffic crash victims. The prompt provision of emergency 
medical care to traffic crash victims will reduce morbidity and 
mortality.
    We believe ACN systems have great potential for reducing deaths and 
injuries caused by motor vehicle accidents. This potential arises from 
the ability of the EDR and ACN, working in tandem, to determine (prior 
to responding to the accident scene) the likely nature and severity of 
the injuries, the proper allocation of resources to respond to those 
injuries, and the location of the crash. We fully expect ACN systems to 
evolve, and our rulemaking today, which standardizes EDR data, will 
play a role in realizing the safety benefits of ACN.
13. Definitions
a. ``Trigger Threshold''
    ``Trigger threshold'' indicates the point at which a recordable 
event is recognized by the EDR as suitable for further analysis. Our 
proposal defined ``trigger threshold'' as ``a change in vehicle 
velocity, in the longitudinal direction for vehicles with only 
longitudinal acceleration measurements or in the horizontal plane for 
vehicles with both longitudinal and lateral measurements, that equals 
or exceeds 0.8 km/h within a 20 ms interval.'' In proposing a value for 
the EDR trigger threshold, we turned to SAE J1698 for guidance.
    GM commented that, as proposed, the trigger threshold for EDR 
recording was set too low and would result in an excessive number of 
recordings and re-recordings. GM argued that the defined threshold 
would create a risk of memory degradation in the electronic control 
module over the life of the vehicle. Accordingly, GM, along with 
DaimlerChrysler, Ford, and Toyota, recommended revising the definition 
of ``trigger threshold'' to read: ``equals or exceeds 5 mph (8 km/h) 
within a 0.15 second interval.'' GM stated that its recommended value 
is consistent with the FMVSS bumper standard threshold.
    Similarly, Hyundai, Kia, and Delphi stated that the trigger 
threshold specified in the NPRM is set too low and would result in data 
being rewritten many times as a result of potholes and curb hits. 
According to the commenters, this frequent overwriting of the EDR data 
could result in computer memory failure, thereby leaving the EDR 
unavailable in the event of an actual crash. Delphi recommended that 
the trigger threshold ``corresponds to an average acceleration in 
excess of 1.5 G with a total velocity change of at least 5 km/hr.''
    As an alternative to the proposed language, TRW Automotive 
suggested that the trigger threshold should be determined by the air 
bag system, which would notify the other systems to begin recording. 
TRW argued that, currently, each individual system records its own data 
so minimal changes would be needed to implement the rule. TRW's 
rationale is that implementation of the rule would be less expensive 
and less complex if the rule permitted each system to record its own 
data.
    TRW Automotive also commented that there should be ``an acceptable 
tolerance of plus or minus ``one data sample period'' for the data 
points corresponding to ``trigger threshold'' detection, and a sampling 
rate tolerance of plus or minus three percent for data before and after 
the point of ``trigger threshold'' detection.''
    Advocates stated that it had no opinion on the exact specification 
for the trigger threshold but expressed concern about setting the 
trigger at a level where recording would occur only in the event of a 
crash. Advocates suggested that NHTSA should consider the collection of 
near-miss data in a future EDR rulemaking. Advocates also questioned 
whether an electrical or engine fire would be a triggering event and 
suggested that NHTSA should revise the rule to require the EDR to be 
sensitive to fire-based events.
    After considering these comments, we have decided to modify the 
trigger threshold value to 8 km/h within a 150 ms interval, as 
requested by the commenters, such that the final rule's definition of 
``trigger threshold'' reads: ``a change in velocity, in the 
longitudinal direction, that equals or exceeds 8 km/h within a 150 ms 
interval. For vehicles that record ``delta-V, lateral,'' trigger 
threshold means a change in vehicle velocity, in either the 
longitudinal or lateral direction that equals or exceeds 8 km/h within 
a 150 ms interval.'' We believe that this change is appropriate for the 
following reasons.
    While we agree that the threshold proposed in the NPRM routinely 
could be exceeded by strong bumper-to-bumper contact in a parking lot 
or minor impact with a road obstacle, we only required the data to be 
recorded if the cumulative delta-V of the current event/crash exceeded 
the delta-V of the previously-recorded data. We do not agree that the 
non-volatile memory would have been over-burdened, because the delta-V 
of the event in non-volatile memory would have rapidly reached a 
sufficient magnitude to disregard minor impacts, such as bumper-to-
bumper events. We believe that the revised criterion effectively 
addresses the concerns raised by the commenters and reduces the 
complications of decisionmaking regarding EDR data recording, while 
maintaining the ability to obtain data from most significant crashes 
(i.e., those that are non-trivial).
    We have decided not to adopt TRW's recommendation to tie the 
trigger threshold to air bag deployment. We are interested in 
collecting data on high delta-V crashes that do not deploy the air bag 
systems. While air bag systems may be operating properly in these

[[Page 51033]]

cases, we are nonetheless interested in these situations, and EDR data 
captured in these situations would be helpful for safety equipment 
analysis. We are also interested in collecting data in non-air bag 
deployment crashes. Finally, one of our stated reasons for this 
rulemaking is to standardize EDRs. We believe that using a set delta-V 
will better facilitate this purpose, whereas using air bag triggers 
could result in different thresholds, depending on manufacturer 
deployment strategies and vehicle platforms. For these reasons, we have 
decided not to narrow our definition of ``trigger threshold'' by tying 
it to air bag deployment.
    Regarding Advocates'' comments recommending capture of near-miss 
data, we have decided that this rulemaking should target crash event 
data. While the agency believes valuable information for crash 
avoidance can be obtained from studying near-miss data, we do not 
believe that current EDRs are best suited for this function. Typically, 
near-miss data are not associated with a strong physical occurrence, 
hence increasing the difficulty of defining a trigger threshold to key 
recording. If the trigger threshold were set very low, it would cause 
the generation of a large volume of files that would need to be 
captured and recorded, or alternatively, it would force EDRs to 
continuously record information. Either of these data logging processes 
would make EDRs much more expensive. At this time, the agency believes 
these issues can be addressed best through our research programs, such 
as the recently completed 100-car study, in which naturalistic driving 
characteristics were captured.\53\ Furthermore, near-miss situations 
are not expected to generate data applicable to the data elements 
selected as non-trivial events in this final rule (e.g., no delta-V or 
safety restraint data).
---------------------------------------------------------------------------

    \53\ Naturalistic Driving Study; Virginia Tech Transportation 
Institute (VTTI); see http://www-nrd.nhtsa.dot.gov/departments/nrd-13/driver-distraction/PDF/100CarMain.pdf.
---------------------------------------------------------------------------

    As with near-miss data, NHTSA does not believe that a trigger 
related to fire would be a cost-effective or practicable approach. Non-
crash fires are typically associated with fuel leaks, and as with the 
near-miss data, current event'driven EDRs would not capture much data, 
even if the EDR were triggered.
b. ``Event''
    In addition to ``trigger threshold,'' the definition of ``event'' 
is important to understanding what constitutes a recordable event for 
an EDR. In the NPRM, we defined ``event'' as ``a crash or other 
physical occurrence that causes the trigger threshold to be met or 
exceeded after the end of the 500 ms period for recording data 
regarding the immediately previous event.''
    GM urged modification of the NPRM's definition of ``event,'' 
arguing that the proposed sampling rates and durations are excessive. 
In order to address these concerns, GM provided a revised definition of 
``event'' and suggested a new definition of ``crash event,'' which also 
sought to clarify the distinction between an event that triggers data 
capture in volatile memory and an event that triggers the recording of 
data in non-volatile memory. DaimlerChrysler, Ford and Toyota offered 
nearly identical comments to those of GM, except that in their 
definition of ``crash event,'' the longitudinal or lateral trigger 
threshold was 5 mph delta-V in 150 ms, as opposed to 5 mph delta-V in 
250 ms for GM.
    Nissan suggested that the rule should permit two alternatives for 
determining the beginning of an event, as provided in SAE J1698. SAE 
J1698 and SAE J1698-1 include two methods of establishing time zero. 
One method calculates time zero as the occurrence of a delta-V of over 
0.8 km/h (0.5 mph) in 20 ms. The other method of calculating delta-V is 
to define time zero as the point at which the EDR algorithm is 
activated, also known as ``wake-up.'' The first method was the basis 
for our proposal in this area. GM, Ford, DaimlerChrysler, and Toyota 
commented that we should first define an ``event'' and then define 
``time zero'' as the beginning of the event, recommending a definition 
of ``event'' as a delta-V of over 8 km/h (5 mph) or more within 150 ms, 
instead delta-V of over 0.8 km/h (0.5 mph) in 20 ms.
    After considering the comments we received on this definition, we 
have slightly modified the definition of ``event'' in the final rule to 
read as: ``a crash or other physical occurrence that causes the trigger 
threshold to be met or exceeded.'' We believe this change is consistent 
with vehicle manufacturers' comments. Under the new trigger threshold 
definition, an event is a physical occurrence that produces sufficient 
delta-V to exceed the trigger threshold. Those occurrences that do not 
meet the threshold are not classified as ``events.''
    As discussed below, we have modified the way in which the start of 
an event and end of an event are determined, consistent with SAE J1698.
c. ``Event Data Recorder''
    The NPRM defined ``event data recorder'' as ``a device or function 
in a vehicle that records any vehicle or occupant-based data just prior 
to or during a crash, such that the data can be retrieved after the 
crash. For purposes of this definition, vehicle or occupant-based data 
include any of the data elements listed in Table I of this part.''
    GM, Ford and Toyota recommended revising the NPRM's definition of 
``event data recorder'' in order to narrow the definition and make it 
more precise.\54\ GM argued that its recommended definition of ``event 
recorder'' would prevent confusion and possible misinterpretation. 
DaimlerChrysler recommended a similar definition for ``event data 
recorder,'' except that DaimlerChrysler's comments omitted the specific 
time references indicated by GM.
---------------------------------------------------------------------------

    \54\ GM offered the following definition of ``event data 
recorder'':
    Event data recorder (EDR) means a device or function in a 
vehicle that captures the data elements identified in Table I of 
this standard for up to 5 seconds before time zero and up to 250 ms 
after time zero, and that records the data when it has been 
determined that a crash event has occurred so that it can be 
retrieved after the crash.
---------------------------------------------------------------------------

    Injury Sciences suggested expanding the definition of EDR to 
include vehicles that record and store any form of speed or collision 
information, without regard to the storage location or purpose. 
According to Injury Sciences, this would prevent manufacturers from 
circumventing the rule by not storing or using the data in their air 
bag modules.
    Gelco commented that the definition of ``event data recorder'' in 
the SUPPLEMENTARY INFORMATION section of the NPRM is narrower than the 
definition in Sec. 563.5 of the proposed regulatory text. Gelco argued 
that the definition in Sec. 563.5 would include devices that are 
designed to capture data at lower resolution on an ongoing basis (as 
distinguished from devices that capture detailed data at the time of a 
crash event.) Gelco stated that such devices have valid purposes for 
both owners and users of vehicles, and that encompassing these devices 
within the definition of EDR would unnecessarily restrain their 
development. Gelco recommended narrowing the scope of the rule by 
adopting a definition for ``event data recorder'' that differentiates 
between devices that capture data on an ongoing basis and EDRs.\55\
---------------------------------------------------------------------------

    \55\ Gelco recommended the following definition of event data 
recorder, in order to clarify the scope of existing recorders 
covered by the rule:
    Event data recorder (EDR) means a device or function installed 
in a vehicle as part of its original equipment that records any 
vehicle or occupant-based data just prior to or during a crash, such 
that the data can be retrieved after the crash. For purposes of this 
definition, vehicle or occupant-based data include any of the data 
elements listed in Table I of this part. For purposes of this 
definition, devices or functions which may record one or more of the 
data elements listed in Table I of this part just prior to or during 
a crash but which are not designed for the purpose of collecting and 
storing motor vehicle crash event data or to record vehicle or 
occupant-based data at the recording intervals/times listed in Table 
I of this part shall not be event data recorders.

---------------------------------------------------------------------------

[[Page 51034]]

    AAM stated that the definition of ``event data recorder'' is too 
broad in that it includes components that are not designed primarily 
for recording crash data. For example, some current recording systems 
only record restraint system deployment decisions and timing data. As a 
result, AAM argued that the rule acts as a mandate forcing 
manufacturers to record a great deal more data than their systems are 
currently designed to record. On the same issue, the Alliance offered 
to help NHTSA draft a specification that more clearly delineates the 
devices that they believe should fall within the ambit of the final 
rule.
    After carefully consideration of the comments, we have decided to 
revise the definition of ``event data recorder'' in order to avoid 
possible misinterpretation. As proposed in the NPRM, the definition 
would have covered all devices that record static freeze-frame air bag 
data elements (e.g., ``frontal air bag warning lamp-on/off''), which 
commenters argued would have inadvertently resulted in a mandatory 
rule. Therefore, we have revised the definition to exclude static 
freeze-frame data elements, and by doing so, we avoid a mandatory rule. 
However, our revised definition retains critical data elements 
necessary for restraint performance evaluation, crash reconstruction, 
and better delta-V estimation.
    The final rule defines ``event data recorder'' as ``a device or 
function in a vehicle that records the vehicle's dynamic, time-series 
data during the time period just prior to a crash event (e.g., vehicle 
speed vs. time) or during a crash event (e.g., delta-V vs. time), 
intended for retrieval after the crash event. For the purposes of this 
definition, the event data do not include audio and video data.''
14. Utilization of SAE and IEEE Standards
    Under Section 563.4, the NPRM proposed to incorporate by reference 
SAE Recommended Practice J211-1, March 1995, ``Instrumentation for 
Impact Test--Part 1--Electronic Instrumentation'' (SAE J211-1). GM 
commented that the proposed section which would have incorporated SAE 
J211-1 should be deleted, arguing that high-speed acceleration data is 
not needed for accident reconstruction purposes (delta-V is sufficient) 
and that manufacturers should have the flexibility to work with their 
suppliers to match data acquisition hardware and software for their 
systems. On the other hand, IEEE-VTS commented that NHTSA should 
include in Section 563.4 several provisions of its consensus Motor 
Vehicle Electronic Data Recorder (MVEDR) standard on a broad range of 
topics.\56\
---------------------------------------------------------------------------

    \56\ IEEE-VTS requested incorporation of the following sections 
of their consensus MVEDR, IEEE 1616 standards: Data Privacy and 
Security Recommendations (Clause 1.3), Definitions (Clause 3.1), 
International Use of MVEDR Data (Clause 4.2), Emergency Response 
Community (i.e. Data Accessibility & Extraction) (Clause 4.3.4), 
Electronic Equipment Operating Environment (Clause 4.6.1), Battery/
Reserve Power (Clause 4.6.2), Crashworthiness (Clause 4.7), Vehicle 
Crash Modes (Clause 5.1), Minimum Outputs (Clause 6.1), Ability to 
Access Nonvolatile Memory (Clause 6.6.), Use of Proprietary 
Connectors (Clause 6.8), MVEDR Telltake (Clause 6.12), Data Capture 
(Clause 7.7), MVEDR Data Dictionary (Clause 8), and Recommended Data 
Elements for Light Vehicles Under 4,500 kg (Clause 8.2).
---------------------------------------------------------------------------

    We have considered GM's comment that Section 563.4 should be 
deleted, which is premised upon replacement of the proposed 
acceleration data element with a delta-V data element. As indicated 
above, manufacturers who prefer to record acceleration may continue to 
do so under this final rule. However, for those manufacturers that 
prefer to record acceleration data instead of delta-V, the acceleration 
data must be filtered and converted to delta-V either during the 
recording period or in the data downloading process. Accordingly, the 
incorporation by reference provision, as it appeared in the NPRM, 
remains relevant, and we see no reason to remove it. We note that the 
incorporated SAE standard is not relevant to manufacturers that decide 
to record delta-V instead of acceleration.
    We have also considered IEEE-VTS's request to incorporate its IEEE 
1616 standard. We note that although incorporation by reference is a 
common practice in our rulemaking, we only utilize it when we believe 
the standards are appropriate and the standards are too complex and 
onerous to be copied into the regulation. In the present case, we 
believe that the provisions of the IEEE standard that do not already 
appear in our proposed EDR rulemaking are not necessary for data 
standardization. For many of the other IEEE provisions that do appear 
in the EDR regulatory text, we do not believe that these standards are 
too complex and onerous to be copied into the regulation. We believe 
that many of the definitions that we have provided in the regulatory 
text are easy to understand and follow. In fact, we believe that it 
would be easier for the reader to understand if all the items were 
articulated in the regulation itself, rather than by incorporation. 
Accordingly, we have we have decided not to incorporate by reference 
the IEEE 1616 standard, as recommended by IEEE-VTS.
15. Costs
    The NPRM estimated that the added cost to manufacturers for 
implementing the requirements of the EDR proposal would be $0.50 per 
vehicle. Several commenters (GM, DaimlerChrysler, Ford, Toyota, Nissan, 
Subaru, ATA, and AIAM) argued that the NPRM's cost estimate is 
understated. These commenters argued that implementation of the 
proposal would result in significantly higher costs related to 
microprocessing and memory upgrades, computer reprogramming, the 
proposed range, accuracy, and precision requirements, the dynamic 
testing requirements, and air bag sensor upgrades. Several commenters 
provided suggestions on ways to reduce costs, while others discussed 
the effect of costs on installation of EDRs.
    GM commented that additional memory and processing capacity 
required to meet the requirements outlined in the NPRM would greatly 
increase the cost of complying with the proposed rule. According to GM, 
memory storage capacity would need to be expanded beyond that provided 
for current EDRs, and memory cannot be added incrementally, as implied 
in NHTSA's cost estimates (i.e., computer memory is normally available 
in blocks, so the next step up from 64K may be 128K). GM further stated 
that microprocessors available to handle larger amounts of memory are 
usually packaged with other system capabilities (e.g., increased input/
output/pins) that would further increase system costs. According to GM, 
this is true for both volatile and non-volatile memory.
    We infer from GM's comments that it believes that, if adopted, our 
proposal would entail unavoidable increases in processor costs. 
Specifically, unless the processor has sufficient memory capacity, the 
ability of the restraint system modules to perform their primary task 
(i.e., deploying the air bags in a timely and appropriate manner) could 
be compromised. GM stated that two microprocessors may be necessary to 
perform these two functions.
    DaimlerChrysler, Ford, and Toyota provided nearly identical 
comments to those of GM on the cost issues associated with memory 
capacity and microprocessing. However, they

[[Page 51035]]

estimated that the NPRM's proposed requirements would necessitate EDR 
storage capacity 5-10 times greater than that found in current EDRs and 
that the overall cost per vehicle would be 2-3 orders of magnitude 
greater than the NPRM's current estimate (i.e., $50-$500). 
DaimlerChrysler and Toyota also argued that costs for RAM memory are 
typically more expensive than ROM memory.
    DaimlerChrysler, Ford, and Toyota commented that the Preliminary 
Regulatory Evaluation's projection of $10,000 per manufacturer for 
software algorithm reprogramming costs is an underestimate, although no 
alternative figure was provided. These manufacturers asserted that such 
efforts would require engineering-level specification development, 
algorithm development, and algorithm validation for each vehicle 
development program.
    GM and AIAM commented that the proposed range, accuracy, and 
precision requirements in Table III of the NPRM underestimate certain 
hardware costs. For example, GM stated that it currently uses  50 G accelerometers with an 8% accuracy. According to GM, 
doubling the range to  100 G and increasing the accuracy of 
those accelerometers would add significant costs, which are not 
reflected in the NPRM's cost estimates. GM added that in some cases, 
the new requirements are beyond the state-of-the-art and may not be 
feasible. AIAM commented that the NPRM specifies range, accuracy and 
precision standards in excess of current industry practice. According 
to these commenters, significant increases in cost would be required to 
modify systems to meet these proposed requirements.
    Another cost issue, raised by GM, Ford, and Toyota, related to the 
proposed dynamic testing requirements for EDRs, which the commenters 
asserted would greatly increase testing costs. For example, GM argued 
that the NPRM would require storage of crashed vehicles for 30 days 
following a test to ensure retrievability of data. GM commented that 
such a requirement is impractical and unnecessary. Ford and Toyota 
challenged the Preliminary Regulatory Evaluation's assumption that the 
NPRM's proposed functionality and survivability requirements would not 
add any costs because existing EDRs are already capable of meeting the 
proposed standard. Ford stated that NHTSA has not fully accounted for 
the crash test performance and survivability provisions, so additional 
costs would be expected.
    As discussed earlier, GM, DaimlerChrysler, and Ford all argued that 
the proposal would significantly increase testing costs, because they 
perceived that the testing would need to be conducted with running 
vehicles and activated systems. According to GM, the NPRM does not 
account for a significant additional cost for reserve or backup 
batteries, which it argued would be necessary to comply with the 
proposed requirement that EDR data be retrievable without external 
power for up to 30 days.
    To remedy the above cost issues, GM recommended reducing the number 
of data elements to only those necessary to obtain safety-related data 
suitable for crash reconstruction purposes, which would presumably 
allow current EDRs to handle these tasks with minimal modifications and 
cost increases.
    Nissan argued that the broad definition of an ``event data 
recorder,'' as proposed, encompasses many current air bag systems that 
do not record the types of information included in Table I. According 
to Nissan's calculations, the NPRM underestimates the cost of 
implementation by a factor of 10. Nissan argued that its air bag 
systems would need major architectural changes to meet the proposed 
requirements. Subaru made a similar comment, arguing that the NPRM 
underestimates the costs of implementation because Subaru might be 
forced to develop an entirely new air bag electronic control unit. AIAM 
commented that some EDR systems that currently only record air bag 
information may need a complete redesign.
    DaimlerChrysler, Ford and Toyota stated that sensors that could 
meet the requirements of the NPRM are currently considered ``laboratory 
grade,'' which raises issues related both to cost and availability.
    Delphi and Mr. Funderburg expressed concern that the cost of 
implementation would deter manufacturers from installing EDRs or take 
away resources from NHTSA's other projects. Delphi commented that the 
cost of implementation might vary significantly depending on the 
existing system architecture and that because of potentially high 
costs, many manufacturers may choose to freeze their level of EDR fleet 
penetration or even remove EDRs from certain models. Commenters argued 
that manufacturers of vehicles with components that are not 
sufficiently interconnected either would remove (or not implement) EDRs 
or would be required to make significant changes to the existing 
electrical architecture. Mr. Funderburg expressed concern regarding the 
costs of data analysis and the potential for diverting NHTSA's 
resources away from more important projects.
    AAA recommended adoption of a smaller data set to help reduce the 
costs of implementation. In contrast, Public Citizen asserted that 
requiring installation of EDRs with an appropriately large number of 
data elements would be more cost-effective for both manufacturers and 
consumers. Public Citizen stated that mandated safety features costs 
consumers as little as a quarter of the cost of such features in the 
absence of an agency requirement. However, Public Citizen did not 
provide any data to substantiate this point.
    We have considered the comments on costs, and we have addressed the 
concerns of the commenters in the Final Regulatory Evaluation (FRE), 
which may be found under the same docket number as this final rule. 
However, the following summarizes the conclusions presented in the FEA.
    The total cost for the estimated 9.8 million vehicles that already 
have an EDR function to comply with the regulation will range up to 
$1.7 million. If manufacturers were to provide EDRs in all 15.5 million 
light vehicles, the estimated total cost will range up to $10.9 
million. These potential costs include technology costs, administrative 
costs, and compliance costs (although the latter two sets of costs are 
expected to be negligible).
16. Other Issues
a. Scope and Purpose
    The NPRM's regulatory text defined the purpose and scope of this 
rulemaking as specifying uniform, national requirements for vehicles 
equipped with EDRs. Proposed section 563.1 also required vehicle 
manufacturers to make sufficient information publicly available to 
enable crash investigators and researchers to retrieve data from EDRs.
    Two vehicle manufacturers commented on the proposed scope 
provision. GM commented that the NPRM's statement of scope is overly 
broad and somewhat ambiguous. GM argued that the current text of Sec. 
563.1 should be revised to clarify the intended scope of the 
regulation, and GM further argued that NHTSA should mandate 
installation of EDRs. Toyota also commented that the scope of the rule 
is overly broad and ambiguous and recommended language nearly identical 
to GM's, but without advocating a mandatory EDR requirement.
    PCIAA commented that the proposed rule focuses too much on 
restraint systems and not enough on systems to help the driver avoid 
collisions.
    We have carefully considered the comments pertaining to the scope

[[Page 51036]]

provision. We disagree with the commenters who stated that our scope 
provision is overbroad and ambiguous. To reiterate our earlier 
explanation, we intend to collect EDR data in order to gather 
information related to crash reconstruction, to the analysis of safety 
equipment performance, and which may be useful for ACN. We believe that 
the regulatory text, when read in its totality (including sections on 
scope, purpose, and definitions), provides the public with a clear 
understanding of the objectives of our final rule.
    We also disagree with commenters' recommendations to change the 
scope of the final rule to adopt a mandatory EDR requirement. As noted 
above, we did not propose a mandatory requirement for vehicle 
manufacturers to install EDRs, and for the reasons previously 
discussed, we have decided not to adopt such an approach at this time. 
We will continue to monitor EDR installation, and may reconsider this 
issue in the future if circumstances warrant. We agree that it is 
desirable for EDRs to gain wider usage and acceptance.
    We have considered PCIAA's comment that the rulemaking should 
acknowledge other uses of EDR data (other than those specified in the 
NPRM) so that data elements offer sufficient flexibility and the 
correct incentives to avoid discouraging innovations that go beyond the 
goals of research and vehicle safety. However, we do not believe that 
this rule will deter EDR innovations beyond NHTSA's stated purposes, 
nor inhibit the ability to use EDRs for other purposes. Furthermore, we 
do not believe it is appropriate to incorporate into this rule other 
uses of EDR data that we currently have no reason to standardize, and 
doing so would require the agency to significantly alter the scope and 
purpose of this rule.
    We have, however, revised the regulatory text of the scope 
provision to make it consistent with the revisions made to the data 
retrieval section. As stated above, in the final rule we have revised 
the portion of our proposal that would have required manufacturers make 
publicly available through the NHTSA docket such non-proprietary 
information that would permit companies that manufacture diagnostic 
tools to develop and build a device capable of accessing, retrieving, 
interpreting, and converting data stored in the EDR. Consistent with 
our new approach arising out of public comments, the scope provision 
now indicates that manufacturers are required under this final rule to 
make such information commercially available.
    DaimlerChrysler recommended adding a time element to the 
``purpose'' section of the regulatory text, stating that EDR recording 
will include ``five seconds of specified pre-crash data elements and 
250 milliseconds of specified crash data elements * * *.'' We have 
considered DaimlerChrysler's recommendation; however, we generally do 
not provide such specific language in the purpose section. Instead, we 
believe that such time element is sufficiently and clearly addressed in 
the regulatory text under the ``data capture'' section.
b. Technical Changes to Definitions and New Definitions
    In response to recommendations provided in the comments, we have 
decided to modify several definitions in the regulatory text. These 
modifications to the regulatory text provide clarification and address 
technical or minor issues.
``Capture''
    The NPRM defined ``capture'' as ``the process of saving recorded 
data.'' GM, DaimlerChrysler, and Ford commented that this definition 
should be clarified. According to GM, the industry defines ``capture'' 
as the process of buffering data in a temporary, volatile storage 
medium where it is continuously updated. GM stated that data captured 
in volatile memory is unstable, insofar as it is continuously 
overwritten with new data as long as power is supplied to the module 
and is lost the moment power is discontinued. We have revised the 
definition of ``capture'' in light of these comments. Accordingly, the 
final rule defines ``capture'' as ``the process of buffering EDR data 
in a temporary, volatile storage medium where it is continuously 
updated at regular time intervals.'' We believe that, as modified, the 
definition of ``capture'' better reflects the industry's understanding 
and uses of that term.
``Record''
    The NPRM defined ``record'' as ``the process of storing data into 
volatile memory for later use.'' GM, DaimlerChrysler, Ford, and Toyota 
recommended changing the definition of ``record'' to ``the process of 
saving captured EDR data into a non-volatile memory storage device for 
subsequent retrieval.'' GM stated that the industry generally uses the 
term ``record'' to mean saving captured data into a non-volatile memory 
storage device that is permanent and stable, even if power is lost to 
the storage module. We agree with these comments and have modified the 
definition of the term ``record'' accordingly. The definition of 
``record'' now reads: ``the process of saving captured EDR data into a 
non-volatile device for subsequent retrieval.''
``Engine Throttle, Percent Full'' and ``Service Brake, On and Off''
    The NPRM defined ``engine throttle, percent full'' as ``for 
vehicles powered by internal combustion engines, the percent of the 
engine throttle opening compared to the full open position of the 
engine throttle opening, and for vehicles not powered by internal 
combustion engines, the percent of vehicle accelerator depression 
compared to the fully depressed position.'' The NPRM defined ``service 
brake, on, off'' as ``the vehicle's service brake is being applied or 
not being applied.''
    GM, DaimlerChrysler, Ford, Toyota, and AIAM recommended revising 
the definition of ``engine throttle, percent full'' to clarify that it 
is the driver input that is recorded, rather than the electrical or 
mechanical output that resulted. The commenters recommended the same 
type of change for the definition of ``service brake, on, off.'' GM's 
rationale is that, while the input and output signals will generally 
correspond, the former is more relevant for safety-related crash 
analyses. AIAM commented that the ``engine throttle, percent full'' 
data element should be redefined to allow the recording of the throttle 
pedal input angle as an alternative means of capturing driver behavior.
    After consideration of these comments, we have determined that both 
definitions should be clarified, as suggested, to reflect that it is 
the driver input that is to be recorded. As stated above in our 
discussion regarding the ``engine RPM'' data element, we believe that 
driver input is more useful for studying crash reconstruction. 
Therefore, the definition of ``engine throttle, percent full'' has been 
clarified and now reads: ``the driver requested acceleration as 
measured by the throttle position sensor on the accelerator pedal 
compared to the fully depressed position.''
    In the final rule, we have also applied this rationale to the 
definition of ``service brake, on/off'' as suggested by the public 
comments, clarifying that it is the driver input that is recorded. The 
new definition reads, ``the status of the device that is installed in, 
or connected to, the brake pedal system to detect whether the pedal was 
pressed. The device can include the brake pedal switch or other driver-
operated service brake control.'' We believe that this definition is 
more suitable for the stated purposes of this rulemaking.

[[Page 51037]]

``Frontal Air Bag''
    The NPRM defined ``frontal air bag'' as ``the primary inflatable 
occupant restraint device that is designed to deploy in a frontal crash 
to protect the front seat occupants.'' GM, DaimlerChrysler, Ford, and 
Toyota recommended revising the NPRM's definition of ``frontal air 
bag'' to make it more closely align to the language of FMVSS No. 
208.\57\ We agree with the commenters and have made this modification 
in the final rule.
---------------------------------------------------------------------------

    \57\ Specifically, GM recommended the following definition:
    Frontal air bag means any inflatable restraint system that 
requires no action by vehicle occupants and is used to meet the 
applicable frontal crash protection requirements of S5.1.2(b) of 
FMVSS No. 208.
---------------------------------------------------------------------------

``Ignition Cycle, Crash'' and ``Ignition Cycle, Download''
    In defining the terms ``ignition cycle, crash'' and ``ignition 
cycle, download,'' the NPRM used the phrase ``ignition key 
applications.'' GM, DaimlerChrysler, Ford, and Toyota recommended 
revising these definitions to reflect that in the future, technological 
changes may obviate the need for a conventional ignition key.
    Based upon these comments, we have modified the relevant 
definitions in the final rule as follows: ``ignition cycle, crash'' 
means ``the number of power cycles applied to the recording device up 
to and including the time when the crash event occurred since the first 
use of the EDR.'' ``Ignition cycle, download'' means ``the number of 
power cycles applied to the recording device at the time when the data 
was downloaded since the first use of the EDR.''
``Normal Acceleration''
    The NPRM defined ``normal acceleration'' as ``the component of the 
vector acceleration of a point in the vehicle in the z-direction. The 
normal acceleration is positive in a downward direction.'' Delphi 
recommended that NHTSA provide greater specificity in the definition of 
0 G normal acceleration, because the term 0 G is used inconsistently 
within the industry (e.g., 0 G is sometimes normalized for the ``1 G 
bias due to gravity). We agree with Delphi's comments and have revised 
the definition. Since the acceleration data are used to compute 
velocity and motion relative to the other vehicle/barrier in our 
laboratory tests, 0 G vertical is defined with the gravity term not 
removed, hence 0 G vertical would be observed when the vertical 
accelerometer is as rest.
``Pretensioner''
    The NPRM defined ``pretensioner'' as ``a device that is activated 
by a vehicle's crash sensing system and removes slack from a vehicle 
belt system.'' GM, DaimlerChrysler, Ford, and Toyota requested a minor 
change in the definition of the term ``pretensioner'' to clarify that 
vehicle belt system means vehicle safety belt system. We agree that the 
addition of the word ``safety'' provides clarity, and we have revised 
the term.
``Safety Belt Status''
    The NPRM defined ``safety belt status'' as ``an occupant's safety 
belt is buckled or not buckled.'' GM, DaimlerChrysler, Ford, and Toyota 
recommended modifying the term to read: ``safety belt status means the 
feedback, as recorded by the EDR function, from the safety system that 
is used to determine that the safety belt is fastened.'' The 
commenters' rationale is that some safety belt technologies provide 
safety belt status feedback without evaluation of the buckle status. We 
agree and have modified the definition in accordance with the 
recommendations. The definitions for both driver and right front 
passenger ``safety belt status'' now read: ``the feedback from the 
safety system that is used to determine that an occupant's safety belt 
is fastened or not fastened.''
``Side Air Bag'' and ``Side Curtain/Tube Air Bag''
    The NPRM defined ``side air bag'' as ``any inflatable occupant 
restraint device that is mounted to the seat or side structure of the 
vehicle interior at or below the window sill, and that is designed to 
deploy and protect the occupants in a side impact crash.'' The proposal 
defined ``side curtain/tube air bag'' as ``any inflatable occupant 
restraint device that is mounted to the side structure of the vehicle 
interior above the window sill, and that is designed to deploy and 
protect the occupants in a side impact crash or rollover.''
    GM and DaimlerChrysler recommended revising the NPRM's definitions 
of ``side air bag'' and ``side curtain/tube air bag'' to simplify the 
locational references in these definitions. GM's recommended 
definitions would also drop the phrase ``and that is designed to deploy 
and protect the occupants in a side impact crash,'' as it appears in 
the NPRM. GM's rationale is that the agency's current definitions do 
not fully comprehend evolving technology that may permit side curtains 
in a variety of locations. Ford provided a nearly identical comment. 
However, Ford recommended adding that these devices are ``designed to 
help mitigate occupant injury and/or ejection.''
    After considering the comments by GM, DaimlerChrysler, and Ford, we 
have modified our definitions of ``side air bag'' and ``side curtain/
tube air bag'' to provide more flexibility for evolving technology. 
However, we do believe that consumers need to know the conditions under 
which side air bags will deploy. To that end, we have deleted the 
specific mounting location references (i.e., above the window sill) 
from the definitions and accepted Ford's recommendation, but retained 
the language that the devices will deploy ``in a side impact'' crash 
event.
    In the final rule, the definition of ``side air bag'' now reads as 
``any inflatable occupant restraint device that is mounted to the seat 
or side structure of the vehicle interior, and that is designed to 
deploy in a side impact crash to help mitigate occupant injury and/or 
ejection.'' The final rule defines ``side curtain/tube air bag'' as 
``any inflatable occupant restraint device that is mounted to the side 
structure of the vehicle interior, and that is designed to deploy in a 
side impact crash or rollover and to help mitigate occupant injury and/
or ejection.''
``Speed, Vehicle Indicated''
    In the NPRM, we proposed to define ``speed, vehicle indicated'' as 
``the speed indicated on the vehicle's speedometer.'' GM, 
DaimlerChrysler, Ford, and Toyota recommended what they believe is a 
more technically correct definition of the ``speed, vehicle 
indicated,'' to read as follows: ``the speed indicated by a 
manufacturer-designated subsystem designed to indicate the vehicle's 
ground travel speed during vehicle operation, as recorded by the EDR.'' 
GM's rationale is that there are no data on the vehicle databus that 
indicate the speed actually being displayed to the driver via the 
speedometer. According to GM, vehicle speed should be reported as 
determined by the appropriate vehicle subsystem(s), which vary among 
manufacturers (e.g., wheel speed sensors, driveline shaft sensors, 
differential sensors, or transmission sensors). Nissan commented that 
manufacturers should have the option of recording the vehicle speed 
from a variety of systems (e.g., ABS) instead of the instrument panel 
speed. AIAM provided a similar comment.
    We agree that the definition of ``speed, vehicle indicated'' in the 
final rule should be modified in a matter

[[Page 51038]]

consistent with these recommendations. Accordingly, the definition of 
``speed, vehicle indicated'' now reads: ``the speed indicated by a 
manufacturer-designated subsystem designed to indicate the vehicle's 
ground travel speed during vehicle operation.''
Timing Issues
    GM, DaimlerChrysler, Ford, and Toyota recommended revising the 
NPRM's definitions for ``time to deploy,'' ``time to first stage,'' and 
``time to nth stage'' to clarify that the elapsed time is in 
milliseconds and that those times are to be measured from time zero to 
the time of the air bag deployment command (rather than to the time of 
air bag inflation or air bag firing).
    We agree with the commenters' suggestions for clarification of the 
time data elements for the air bag systems and other commanded systems, 
such as pretensioners. Accordingly, we have revised all relevant 
definitions, including ``time to deploy, pretensioner,'' to reflect 
that these elements are measured to the time of the deployment command 
signal that is generated within the control unit.
``Time Zero'' and ``End of Event Time''
    The NPRM defined ``time zero'' as the ``beginning of the first 20 
ms interval in which the trigger threshold is met during an event.'' 
GM, DaimlerChrysler, Ford, and Toyota recommended revising the 
definition for ``time zero'' in order to better standardize a common 
reference point for all EDR data, thereby facilitating comparisons 
among data sets from different vehicles. GM proposed new language for 
that definition.
    We have reviewed this definition of ``time zero'' in conjunction 
with our newly adopted definition of ``trigger threshold,'' and we have 
taken into account the different types of EDR system algorithms (e.g., 
ones with continuously running algorithms, as opposed to ones using an 
algorithm ``wake-up'' strategy). As discussed above, we have revised 
the definition of ``trigger threshold'' to mean ``8 km/h within a 150 
ms interval.'' This defines the crash level that will be captured and 
recorded in the EDR. We acknowledge that OEMs use different operational 
strategies to sense a crash in their air bag control modules. For 
example, some manufacturers use a continuously operating system that is 
always on and sensing acceleration and analyzing the signal(s) to make 
an air bag command decision. In contrast, other manufacturers utilize 
systems that ``wake up'' when a crash occurs.
    We agree that ``time zero'' needs to be defined so as to ensure 
that each of these strategies will result in similar crash data time 
reporting in the EDR record. To accomplish this, NHTSA has turned to 
SAE J1698 for additional guidance. SAE, working with members from 
companies that employ the two operating strategies, has worked out 
these issues, so we have adopted this approach, as discussed below.
    For systems that wake up, ``time zero'' is defined as the time the 
control algorithm is activated. When a crash occurs, the system wakes 
up almost instantly, and it starts processing the crash data. Thus, 
``time zero'' is established at or very close to the time the crash 
starts. ``Wake up'' is typically determined by the accelerometer 
exceeding a pre-defined threshold for a pre-defined time period, such 
as 2 G for 1 ms. The data are captured, and if the delta-V exceeds 8 
km/h with in a 150 ms interval, the data are recorded.
    For systems with continuously running algorithms, the ``time zero'' 
determination is more complicated. In such systems, the CPU (central 
processing unit) is continuously processing accelerometer data in order 
to make air bag command decisions. SAE decided, for these systems, that 
the start of an event should be defined by a change in velocity. Thus, 
we have adopted the same strategy. For systems that run continuously, 
we are defining ``time zero'' as the first time point where a 
longitudinal, cumulative delta-V of over 0.8 km/h (0.5 mph) is reached 
within a 20 ms time period. Since acceleration rises quickly in a major 
crash, we anticipate that this strategy should work well, resulting in 
time zeros in good agreement with the ``wake up'' systems. Thus, for 
continuously-running systems, the CPU monitors the vehicle's 
deceleration signal(s). If the total delta-V exceeds 8.0 km/h within a 
150 ms period, an event is detected and the captured data are recorded.
    In lateral crashes, the longitudinal trigger may not be triggered, 
and in those cases, there would be no data recorded in the EDR. For 
vehicles that choose to record ``delta-V, lateral,'' we are extending 
the trigger threshold and time zero definition so that in those 
vehicles, EDR data is recorded. We have turned to SAE J1698-1 for the 
time zero definition, selecting time zero as the first point in the 
interval where the cumulative, lateral delta-V equals or exceeds 0.8 
km/h (0.5 mph) within a 5 ms interval.
    To facilitate detection of a second event in a multi-event crash, 
we have added a new definition to automate the detection of the end of 
an event. After once again consulting SAE J1698-1, we have defined 
``end of event time'' as the moment when the cumulative delta-V within 
a 20 ms time period becomes 0.8 km/h (0.5 mph) or less. This will allow 
manufacturers to develop methodologies to automatically detect the end 
of one event and start processing data to determine whether a second 
event occurs during the crash.
New Definitions
    In reviewing our proposal and after making substantive 
modifications to other parts of the final rule based on the public 
comments, we have added a few terms to the ``definitions'' section of 
the final rule's regulatory text for clarification purposes. The new 
terms are: (1) ``Air bag warning lamp status,'' (2) ``deployment time, 
frontal air bag,'' (3) ``volatile memory buffer,'' (4) ``non-volatile 
memory buffer,'' (5) ``occupant position classification,'' and (6) 
``end of event time.'' We also modified the definitions of ``occupant 
size classification'' and ``seat position'' to make them more flexible 
to account for developing technologies.
c. Data Capture
    In the NPRM, we explained that once the trigger threshold has been 
met or exceeded, EDR data elements are captured in volatile memory. We 
further explained that the EDR continues to capture data for an 
additional 500 ms. The EDR makes a determination (by comparing the 
absolute values of the maximum delta-V captured with the data 
previously recorded) of whether to discard the EDR data captured in 
favor of a previously recorded data set. We proposed a specific 
hierarchy on how an EDR should capture and record data, including data 
in cases of multi-event crashes. This strategy was proposed so that the 
EDR would retain crash data associated with the higher maximum delta-V. 
We developed this method in the NPRM to ensure that the EDR does not 
overwrite an important file generated in a crash with a minor 
subsequent event, such as loading a crashed car on a wrecker.
    GM, DaimlerChrysler, Ford, and Toyota recommended that NHSTA delete 
subparagraphs (a), (b), (c), (f) from our proposed regulatory text 
section on ``data capture.'' Those commenters also suggested that NHSTA 
replace subparagraphs (d) and (e), which discuss data capture 
requirements associated with air bag deployment, with the following 
language: ``a non-deployment event will overwrite a non-deployment 
event of lesser magnitude; deployment events must always overwrite non-
deployment events; deployment events must lock the record and may not 
be overwritten.''

[[Page 51039]]

    In their comments, GM, DaimlerChrysler, Ford, and Toyota stated 
that the NPRM's proposed data capture requirements are complex and 
ambiguous and do not accurately recognize the system architecture in 
restraint control modules. These four commenters also stated that the 
requirements do not take into consideration the limitations of current 
technology. They argued that it is impractical to simultaneously buffer 
data, to write data to nonvolatile memory, to analyze the severity of 
the impact(s), and to appropriately deploy restraints.
    To reduce the risk of data loss, Nissan stated that an air bag 
deployment event should be written to memory and locked, and that all 
further recording should cease. Advocates questioned whether a 5-second 
window is sufficient to capture an entire post-crash event.
    We have carefully considered the comments and have developed a 
modified strategy for making the data recording decision, based on the 
comments submitted by the manufacturers. We have adopted these 
commenters' suggestions for a new definition of ``trigger threshold,'' 
and based upon this new definition, all crashes captured and recorded 
will be of significant magnitude to be of interest. Thus, the 
comparative process, as proposed, is no longer necessary.
    We also have decided that collecting data associated with an air 
bag event is our priority. Accordingly, in the final rule, we have 
specified a new capture logic that accounts for the comments, 
simplifies the EDR design, reduces the risk of losing important air bag 
data, and will likely reduce costs.
    The new methodology requires the EDR to make two analytical 
decisions: one is related to an air bag crash event, and the other is 
related to a non-air bag crash event. In those crash events where an 
air bag is commanded to deploy, the EDR must delete the data previously 
recorded, and the data from the air bag crash event must be captured, 
recorded, and locked to prevent overwriting. In those crashes where air 
bags are not commanded to deploy, our logic deletes all previously 
captured and recorded data, for up to two events. If the second event 
turns out to be air bag related, the logic calls for a revision to the 
first condition. In these cases, collection of the first non-air bag 
related event is not necessary but is acceptable. We believe that this 
logic provides relief in terms of the need for increased CPU power that 
might otherwise be necessary for an EDR to analyze and capture EDR data 
during a time when it might complicate safety-critical decisions.
d. Miscellaneous Comments
    SEMA urged NHTSA to refrain from adopting requirements that could 
ossify EDR technology, commenting that the EDR system needs to be 
adaptable to allow for future developments and to work with other 
vehicle systems. According to SEMA, the system should not preclude 
servicing, repair, or installation of aftermarket equipment. SEMA 
argued that manufacturers, distributors, dealers, and motor vehicle 
repair businesses must have sufficient information about the EDR system 
to be able to service the vehicle and to install new or replacement 
products without fear of taking vehicle equipment out of compliance 
with any applicable Federal motor vehicle safety standard. In response 
to SEMA's comments, we do not believe that these systems will be any 
more complicated than current air bag control systems. EDRs are not new 
to the marketplace, and no specific problem of this sort has been 
brought to NHTSA's attention.
    NADA, EPIC, and Honda commented on the need for public education 
and awareness of EDRs. NADA stated that NHTSA should work to educate 
the public ``that, in addition to the potential for improving vehicle 
and roadway safety design effectiveness, appropriately utilized EDR 
system information will help to reduce accident-related investigation, 
medical, legal, and insurance costs.'' EPIC commented that currently, 
public awareness and understanding of EDRs is insufficient. EPIC urged 
NHTSA to create an EDR information website to educate the public about 
EDR technology and its uses, what types of users may gain access to EDR 
information and the circumstances under which it may be accessed, and 
privacy rights associated with EDR data.
    NHTSA agrees with the value of a Web site dedicated to EDRs. About 
five years ago, NHTSA launched the first EDR Web site. The Web site 
contained historical information about EDR technology, research 
material regarding EDR uses, patent information and other resources. In 
late 2004, NHTSA commenced work on a full update to the Web site, which 
was completed in early 2005. It is accessible through NHTSA's Web site, 
http://www.nhtsa.dot.gov, and at http://safercar.gov.
    Concurrent with the publication of this final rule, we are posting 
a consumer-directed set of ``questions and answers'' on our Web site to 
provide educational materials and to raise awareness about the presence 
and functionality of EDRs. Topics include common privacy concerns and 
NHTSA's protocol for requesting EDR data during crash investigations. 
In developing these materials, we will consider NADA's recommendations 
to inform consumers that EDRs could lead to reductions in accident-
related investigation, medical, legal, and insurance costs. Our plan is 
to keep these materials up to date, by adding new information as unique 
questions from the public are raised.
    Honda suggested that NHTSA should conduct an EDR workshop so that 
all critical issues can be explored and discussed, thereby facilitating 
issuance of a final rule in an expedient fashion and minimizing the 
need for petitions for reconsideration. Although an EDR workshop, as 
recommended by Honda, would offer a means of gaining additional EDR-
related input, we have decided that such a meeting is not necessary 
before proceeding to a final rule.
    ATA stated that NHTSA should conduct additional human factors 
research to determine the effect of driver and employee awareness of 
EDRs on the number and severity of crashes. ATA's comment pertains to 
research, not to this final rule. We note, however, that we believe the 
issue of EDR awareness as related to the number and severity of crashes 
may be a valuable area for future research.
    Public Citizen offered additional recommendations, including: (1) 
NHTSA should to fully integrate EDR data into all of its data 
collection systems and crash investigations; (2) police and municipal 
officials should be trained to enable them to collect accurate and 
complete EDR data for the Fatality Analyses Reporting System (FARS) 
database; and (3) NHTSA should create a new database solely for EDR 
data.
    We agree with Public Citizen regarding the value of incorporating 
EDR data into our national databases. Starting in 2000, NHTSA began to 
routinely collect EDR data in our NASS/CDS, SCI, and Crash Injury 
Research and Engineering Network (CIREN) data systems. To date, we have 
colleted over 2,700 cases with EDR data. However, we are not collecting 
EDR data in FARS at this time. The agency is working with police 
officials to develop guidelines for training classes to ensure that EDR 
data are downloaded properly and that these officials are educated on 
the limitations of these devices.
    The European Communities requested that the U.S. refrain from 
finalizing its EDR proposal until there has been an opportunity for 
further consultations both bilaterally and in international fora. The 
European Communities'

[[Page 51040]]

rationale is that EDRs have been identified as an item for bilateral 
research cooperation between NHTSA and the Directorate-General 
Enterprise of the European Commission. The European Communities also 
noted that the World Forum for Harmonization of Vehicle Regulations 
(WP.29), administered by the UN Economic Commission for Europe (ECE), 
has agreed to establish an informal working group on EDRs. The European 
Communities expressed hope that with U.S. participation, it would be 
possible to develop a global technical regulation for EDRs.
    We have carefully considered the EC's comments. NHTSA has concluded 
that it needs to move forward at this time with a basic set of 
requirements, because EDR data can help the government and industry 
better understand crash events and safety system performance, thereby 
contributing to safer vehicle designs and more effective safety 
regulations. EDR data can also play a role in advancing developing 
networks for providing emergency medical services, such as ACN. The 
agency has sought to establish this foundation in a way that would 
encourage broad application of EDR technologies in motor vehicles and 
maximize the usefulness of EDR data for researchers, regulators, and 
the medical community, while avoiding the imposition of unnecessary 
burdens or hampering future improvements to EDRs.
    NHTSA looks forward to continuing work on this issue with the 
European Communities, as well as with the international community under 
the auspices of the World Forum for the Harmonization of Vehicle 
Regulations administered by the United Nations. The action taken today 
in no way precludes achieving common understandings in the future.
    Mr. Bretherton, an individual, commented that better coordination 
of Traffic Records Coordinating Committees (TRCCs) within States is 
needed to facilitate the use of crash data and that funding is needed 
to address technology needs, to make data uniform between States, and 
to ensure data collection by all States. He expressed concern that 
local governments may have increased liability as a result of crash 
data. He also stated that ``Fast FARS'' is not a good use of resources. 
Again, although these issues are worth considering at an appropriate 
time and in an appropriate forum, they are beyond the scope of the 
present rulemaking.

V. Rulemaking Analyses and Notices

A. Vehicle Safety Act

    Under 49 U.S.C. 322(a), the Secretary of Transportation (the 
``Secretary'') has authority to prescribe regulations to carry out 
duties and power of the Secretary. One of the duties of the Secretary 
is to administer the National Traffic and Motor Vehicle Safety Act, as 
amended. The Secretary has delegated the responsibility for carrying 
out the National Traffic and Motor Vehicle Safety Act to NHTSA.\58\
---------------------------------------------------------------------------

    \58\ 49 U.S.C. 105 and 322; delegation of authority at 49 CFR 
1.50.
---------------------------------------------------------------------------

    We note that in 1994, the National Traffic and Motor Vehicle Safety 
Act, as amended, was repealed and simultaneously codified into 49 
U.S.C. Chapter 301, Motor Vehicle Safety, by Pub. L. 103-272 (July 5, 
1994). This involved moving these provisions from 15 U.S.C. Chapter 38 
to 49 U.S.C. Chapter 301. Section 1(a) of Pub. L. 103-272 stated that 
the laws codified were so codified ``without substantive change.'' 
Prior to this codification, a specific provision in 15 U.S.C. 1407 
provided, ``The Secretary is authorized to issue, amend, and revoke 
such rules and regulations as he deems necessary to carry out this 
subchapter.'' However, in the codification process, this provision was 
deleted as unnecessary, because, as specifically noted in the 
legislative history, the Secretary already had such powers pursuant to 
49 U.S.C. 322(a).\59\ Thus, the Secretary, and NHTSA, have general 
authority to issue such rules and regulations as deemed necessary to 
carry out Chapter 301 of Title 49, United States Code.
---------------------------------------------------------------------------

    \59\ H.R. Rep. No. 103-180, Table 2A, at 584 (1993).
---------------------------------------------------------------------------

    Under 49 U.S.C. Chapter 301, Motor Vehicle Safety (49 U.S.C. 30101 
et seq.), the Secretary of Transportation, and, by delegation, NHTSA, 
is responsible for prescribing motor vehicle safety standards that are 
practicable, meet the need for motor vehicle safety, and are stated in 
objective terms.\60\ These motor vehicle safety standards set the 
minimum level of performance for a motor vehicle or motor vehicle 
equipment to be considered safe.\61\ When prescribing such standards, 
NHTSA must consider all relevant, available motor vehicle safety 
information.\62\ NHTSA also must consider whether a proposed standard 
is reasonable, practicable, and appropriate for the type of motor 
vehicle or motor vehicle equipment for which it is prescribed and the 
extent to which the standard will further the statutory purpose of 
reducing traffic accidents and associated deaths.\63\
---------------------------------------------------------------------------

    \60\ 49 U.S.C. 30111(a).
    \61\ 49 U.S.C. 30102(a)(9).
    \62\ 49 U.S.C. 30111(b).
    \63\ Id.
---------------------------------------------------------------------------

    Similar to our approach in the area of vehicle identification 
numbers, we decided to develop a general regulation for EDRs rather 
than a Federal motor vehicle safety standard. We did not believe it was 
appropriate to issue an FMVSS that would trigger the statute's recall 
and remedy provisions, because the benefits of EDRs are expected to be 
derivative from better crash-related information, rather than having a 
direct impact on the safety of the individual vehicle equipped with an 
EDR. A failure to meet the EDR requirements would, however, be subject 
to an enforcement action. While we have not issued the regulation as an 
FMVSS, however, we have generally followed the statutory requirements 
that apply to FMVSSs.
    First, this final rule was preceded by an initial request for 
comments and an NPRM, which facilitated the efforts of the agency to 
obtain and consider relevant motor vehicle safety information, as well 
as public comments. Further, in preparing this document, the agency 
carefully evaluated available research, testing results, and other 
information related to various EDR technologies. We have also updated 
our economic estimates and analyses to account for new cost information 
provided by public commenters. In sum, this document reflects our 
consideration of all relevant, available motor vehicle safety 
information.
    Second, to ensure that the EDR requirements are practicable, the 
agency considered the cost, availability, and suitability of requiring 
various EDR data elements, consistent with our safety objectives. We 
note that EDRs are already installed on most light vehicles, and 
because the data elements in the final rule are to a large extent 
already incorporated in EDRs, we believe that it will be practicable to 
standardize these data elements in light vehicles voluntarily equipped 
with EDRs and that such incremental changes will be minor. In light of 
the steady advances made in EDR technologies over the past few years, 
we believe that vehicle manufacturers will have a number of 
technological choices available for meeting the requirements of the 
final rule for EDRs. In sum, we believe that this final rule is 
practicable and will provide several benefits, including provision of 
better pre-crash and crash-related data that may be valuable for 
designing safer vehicles and for use by medical first responders.

[[Page 51041]]

    Third, the regulatory text following this preamble is stated in 
objective terms in order to specify precisely what performance is 
required and how performance will be tested to ensure compliance with 
the regulation. Specifically, the final rule sets forth performance 
requirements for operation of the EDRs, including the type of data that 
the EDR must capture and record, the data's range/accuracy/resolution, 
and the data's retrievability.
    The final rule also includes test requirements for the 
survivability of EDR data through reference to existing crash test 
requirements in other FMVSSs (i.e., Standard Nos. 208 and 214). This 
approach helps ensure that EDR data survive most crashes without 
establishing news kinds of vehicle tests. The test procedures under 
FMVSS Nos. 208 and 214 already carefully delineate how testing is 
conducted. Thus, the agency believes that these test procedures are 
sufficiently objective and will not result in any uncertainty as to 
whether a given vehicle satisfies the requirements of the EDR 
regulation.
    Fourth, we believe that this final rule will meet the need for 
motor vehicle safety because the EDR regulation will help researchers 
better understand pre-crash and crash events. Standardization of EDR 
data should improve the consistency and comparability of these data. 
This information will be useful to NHTSA, vehicle manufacturers, and 
other interested stakeholders for a variety of purposes, including 
developing safety vehicle designs and more effective regulations. In 
addition, standardized EDR data may be useful for ACN and other systems 
for providing emergency medical services.
    Finally, we believe that this final rule is reasonable and 
appropriate for motor vehicles subject to the applicable requirements 
(i.e., light vehicles voluntarily equipped with EDRs). As discussed 
elsewhere in this notice, the agency has sought to limit the minimum 
data set in this final rule to those elements necessary to achieve the 
agency's stated purposes and to minimize the burdens associated with 
the regulation. We believe that because most EDRs already possess many 
of these capabilities, any required adjustments should be minor. 
Accordingly, we believe that this final rule is appropriate for covered 
vehicles that are or would become subject to these provisions of the 
EDR regulation because it furthers the agency's objective of preventing 
deaths and serious injuries through better understanding of crash-
related events that may lead to safer vehicle designs and more 
effective regulations.

B. Executive Order 12866 and DOT Regulatory Policies and Procedures

    Executive Order 12866, ``Regulatory Planning and Review'' (58 FR 
51735, October 4, 1993), provides for making determinations whether a 
regulatory action is ``significant'' and therefore subject to OMB 
review and to the requirements of the Executive Order. The Order 
defines a ``significant regulatory action'' as one that is likely to 
result in a rule that may:
    (1) Have an annual effect on the economy of $100 million or more or 
adversely affect in a material way the economy, a sector of the 
economy, productivity, competition, jobs, the environment, public 
health or safety, or State, local, or Tribal governments or 
communities;
    (2) Create a serious inconsistency or otherwise interfere with an 
action taken or planned by another agency;
    (3) Materially alter the budgetary impact of entitlements, grants, 
user fees, or loan programs or the rights and obligations of recipients 
thereof; or
    (4) Raise novel legal or policy issues arising out of legal 
mandates, the President's priorities, or the principles set forth in 
the Executive Order.
    This final rule has been determined to be significant, and the 
agency has prepared a separate document, a Final Regulatory Evaluation, 
addressing the benefits and costs for the rule. (A copy is available in 
the docket for this rulemaking.) As a significant notice, it was 
reviewed under Executive Order 12866. The rule is also significant 
within the meaning of the Department of Transportation's Regulatory 
Policies and Procedures. While the potential cost impacts of the final 
rule are far below the level that would make this a significant 
rulemaking, the rulemaking addresses a topic of substantial public 
interest.
    As discussed in that document and in the preceding sections of this 
final rule, the crash data that will be collected by EDRs under this 
rule will be valuable for the improvement of vehicle safety. We believe 
that the EDR data we collect will improve crash investigations, the 
evaluation of safety countermeasures, advanced restraint and safety 
countermeasure research and development, and advanced ACN. However, the 
improvement in vehicle safety will not occur directly from the 
collection of crash data by EDRs, but instead from the ways in which 
the data are used by researchers, vehicle manufacturers, ACN and EMS 
providers, government agencies, and other members of the safety 
community. Therefore, it is not presently practical to quantify the 
safety benefits.
    We estimate that about 64 percent of new light vehicles are already 
equipped with EDRs. As discussed earlier, vehicle manufacturers have 
provided EDRs in their vehicles by adding EDR capability to their 
vehicles' air bag control systems. The costs of EDRs have been 
minimized, because they involve the capture into memory of data that is 
already being processed by the vehicle, and not the much higher costs 
of sensing much of that data in the first place.
    The costs of the rule will be the incremental costs for vehicles 
equipped with EDRs to comply with the requirements. As discussed in the 
agency's separate document on benefits and costs, we estimate the total 
costs of the final rule will range up to $1.7 million. While the 
potential costs include technology costs, administrative costs, and 
compliance costs, the administrative and compliance costs are estimated 
to be negligible. The final rule will not require additional sensors to 
be installed in vehicles, and the primary technology cost will result 
from a need to upgrade EDR memory chips. The total cost for the 
estimated 9.8 million vehicles that already have an EDR function to 
comply with the regulation will range up to $1.7 million. If 
manufacturers were to provide EDRs in all 15.5 million light vehicles, 
the estimated total cost will range up to $10.9 million. A complete 
discussion of how NHTSA arrived at these costs may be found in the 
separate document on benefits and costs.

C. Regulatory Flexibility Act

    NHTSA has considered the impacts of this rulemaking action under 
the Regulatory Flexibility Act (5 U.S.C. 601 et seq.) I certify that 
the final rule will not have a significant economic impact on a 
substantial number of small entities.
    The following is the agency's statement providing the factual basis 
for the certification (5 U.S.C. 605(b)). This rule directly affects 
motor vehicle manufacturers, second stage or final manufacturers, and 
alterers. Business entities are defined as ``small businesses'' using 
the North American Industry Classification System (NAICS) code, for the 
purposes of receiving Small Business Administration assistance. One of 
the criteria for determining size, as stated in 13 CFR 121.201, is the 
number of employees in the firm. Affected business categories include 
the following. To qualify as a small business in: (a) Automotive 
Manufacturing (NAICS 336111), the

[[Page 51042]]

firm must have fewer than 1,000 employees; (b) Light Truck and Utility 
Vehicle Manufacturing (NAICS 336112), the firm must have fewer than 
1,000 employees; (c) Motor Vehicle Body Manufacturing (NAICS 336211), 
the firm must have fewer than 1,000 employees; (d) All Other Motor 
Vehicle Parts Manufacturing (NAICS 336399), the firm must have fewer 
than 750 employees; (e) Computer Storage Manufacturers (NAICS 334111), 
the firm must have fewer than 1,000 employees, and (f) Software 
Reproducing (NAICS 334611), the firm must have fewer than 500 
employees.
    Only four of the 18 motor vehicle manufacturers affected by this 
rule qualify as a small business. Most of the intermediate and final 
stage manufacturers of vehicles built in two or more stages and 
alterers have 1,000 or fewer employees. However, these small businesses 
adhere to original equipment manufacturers' instructions in 
manufacturing modified and altered vehicles. Based on our knowledge, 
original equipment manufacturers do not permit a final stage 
manufacturer or alterer to modify or alter sophisticated devices such 
as air bags or EDRs. Therefore, multistage manufacturers and alterers 
will be able to rely on the certification and information provided by 
the original equipment manufacturer. Accordingly, there will be no 
significant impact on small business, small organizations, or small 
governmental units by these amendments.

D. Executive Order 13132 (Federalism)

    Executive Order 13132 sets forth principles of federalism and the 
related policies of the Federal government. As noted above, NHTSA 
expects that general principles of preemption law would operate so as 
to displace any conflicting State law or regulations (for further 
discussion of preemption, see section IV.B.9 above).
    NHTSA sought comment from all stakeholders on the issue of 
preemption through publication of the proposed rule in the Federal 
Register. NHTSA received one comment on the proposed rule from State 
and local governmental entities.
    Additionally, officials at NHTSA consulted with organizations 
representing the interests of state and local governments and officials 
about this rulemaking and the issue of preemption.
    NHTSA has complied with Executive Order 13132 and has determined 
that this final rule is consistent with its provisions.

E. Executive Order 12988 (Civil Justice Reform)

    Executive Order 12988 requires that agencies review proposed 
regulations and legislation and adhere to the following general 
requirements: (1) The agency's proposed legislation and regulations 
shall be reviewed by the agency to eliminate drafting errors and 
ambiguity; (2) The agency's proposed legislation and regulations shall 
be written to minimize litigation; and (3) The agency's proposed 
legislation and regulations shall provide a clear legal standard for 
affected conduct rather than a general standard, and shall promote 
simplification and burden reduction.
    When promulgating a regulation, Executive Order 12988, specifically 
requires that the agency must make every reasonable effort to ensure 
that the regulation, as appropriate: (1) Specifies in clear language 
the preemptive effect; (2) specifies in clear language the effect on 
existing Federal law or regulation, including all provisions repealed, 
circumscribed, displaced, impaired, or modified; (3) provides a clear 
legal standard for affected conduct rather than a general standard, 
while promoting simplification and burden reduction; (4) specifies in 
clear language the retroactive effect; (5) specifies whether 
administrative proceedings are to be required before parties may file 
suit in court; (6) explicitly or implicitly defines key terms; and (7) 
addresses other important issues affecting clarity and general 
draftsmanship of regulations.
    NHTSA has reviewed this final rule according to the general 
requirements and the specific requirements for regulations set forth in 
Executive Order 12988. The issue of the preemptive effect of this final 
rule was discussed in detail in the section on Executive Order 13132 
(Federalism) immediately above, so rather than repeat those points 
here, we would refer readers to that section for a full discussion. A 
petition for reconsideration or other administrative proceeding is not 
required before parties may file suit in court.

F. Executive Order 13045 (Protection of Children From Health and Safety 
Risks)

    Executive Order 13045, ``Protection of Children from Environmental 
Health and Safety Risks'' (62 FR 19855, April 23, 1997), applies to any 
rule that: (1) Is determined to be ``economically significant'' as 
defined under Executive Order 12866, and (2) concerns an environmental, 
health, or safety risk that the agency has reason to believe may have a 
disproportionate effect on children. If the regulatory action meets 
both criteria, the agency must evaluate the environmental health or 
safety effects of the planned rule on children, and explain why the 
planned regulation is preferable to other potentially effective and 
reasonable feasible alternatives considered by the agency.
    Because the EDR final rule is not an economically significant 
regulatory action under Executive Order 12866 and does not involve 
decisions based upon health and safety risks that disproportionately 
affect children, no further analysis under Executive Order 13045 is 
necessary.

G. Paperwork Reduction Act

    GM DaimlerChrysler, Ford, and Toyota commented that the agency's 
NPRM underestimated the paperwork burden associated with section 
563.12's requirement for filing technical instructions for 
manufacturing download devices for each vehicle model. The NPRM 
estimated those paperwork costs as 20 hours per year per manufacturer. 
GM's rationale is that the proposed requirement to file this 
information 90 days prior to the start of production for each vehicle 
model would require a continuous stream of data filings for the 
multiple vehicle launches that full-line manufacturers have throughout 
the calendar year. According to GM, each filing would involve a 
compilation of the technical data, as well as technical and legal 
review, tasks which would require more than 20 hours of work for each 
vehicle model.
    These concerns have been addressed because we have decided not to 
adopt the proposed provision, so deleting those reporting requirements 
eliminates the paperwork costs that had been associated with this 
rulemaking. Thus, there are not any information collection requirements 
associated with this final rule.

H. National Technology Transfer and Advancement Act

    Section 12(d) of the National Technology Transfer and Advancement 
Act of 1995 (NTTAA), Public Law 104-113 (15 U.S.C. 272) directs the 
agency to evaluate and use voluntary consensus standards in its 
regulatory activities unless doing so would be inconsistent with 
applicable law or is otherwise impractical. Voluntary consensus 
standards are technical standards (e.g., materials specifications, test 
methods, sampling procedures, and business practices) that are 
developed or adopted by voluntary consensus standards bodies, such as 
the Society of Automotive Engineers. The NTTAA

[[Page 51043]]

directs us to provide Congress (through OMB) with explanations when the 
agency decides not to use available and applicable voluntary consensus 
standards. The NTTAA does not apply to symbols.
    There are several consensus standards related to EDRs, most notably 
those standards published by SAE and IEEE. NHTSA has carefully 
considered the consensus standards applicable to EDR data elements. 
Consensus standards for recording time/intervals, data sample rates, 
data retrieval, data reliability, data range, accuracy and precision, 
and EDR crash survivability were evaluated by NHTSA and adopted when 
practicable.
    In this final rule, we have incorporated by reference SAE 
Recommended Practice J211-1, March 1995, ``Instrumentation for Impact 
Test--Part 1--Electronic Instrumentation.'' For those manufacturers 
that prefer to record acceleration data instead of or in addition to 
delta-V, SAE J211-1 provides a standard for filtering the acceleration 
data that are then converted to delta-V either during the recording 
period or in the data downloading process.
    Previously in this notice, NHTSA has explained why other voluntary 
consensus standards were not adopted for certain technical standards 
set forth in this rule. For further analysis of the incorporation of 
consensus standards, please refer to section IV.B.14 above.

I. Unfunded Mandates Reform Act

    Section 202 of the Unfunded Mandates Reform Act of 1995 (UMRA) 
requires Federal agencies to prepare a written assessment of the costs, 
benefits, and other effects of proposed or final rules that include a 
Federal mandate likely to result in the expenditure by State, local, or 
tribal governments, in the aggregate, or by the private sector, of more 
than $ 100 million in any one year (adjusted for inflation with base 
year of 1995). Section 205 of the UMRA generally requires that, before 
promulgating a rule for which a written statement is needed, NHTSA 
identify and consider a reasonable number of regulatory alternatives 
and adopt the least costly, most cost-effective, or least burdensome 
alternative that achieves the objectives of the rule. The provisions of 
section 205 do not apply when they are inconsistent with applicable 
law. Moreover, section 205 allows NHTSA to adopt an alternative other 
than the least costly, most cost-effective, or least burdensome 
alternative if the agency publishes with the final rule an explanation 
why that alternative was not adopted.
    This rule does not impose any unfunded mandates under the Unfunded 
Mandates Reform Act of 1995. The rule does not result in the 
expenditure by State, local, or tribal governments, or the private 
sector, in the aggregate, or more than $118 million annually (2004 
dollars). Thus, this final rule is not subject to the requirements of 
sections 202 and 205 of the UMRA.

J. National Environmental Policy Act

    NHTSA has analyzed this rulemaking action for the purposes of the 
National Environmental Policy Act. The agency has determined that 
implementation of this action will not have any significant impact on 
the quality of the human environment.

K. Regulatory Identifier Number

    The Department of Transportation assigns a regulation identifier 
number (RIN) to each regulatory action listed in the Unified Agenda of 
Federal Regulations. The Regulatory Information Service Center 
publishes the Unified Agenda in April and October of each year. You may 
use the RIN contained in the heading at the beginning of this document 
to find this action in the Unified Agenda.

L. Privacy Act

    Please note that anyone is able to search the electronic form of 
all comments received into any of our dockets by the name of the 
individual submitting the comment (or signing the comment, if submitted 
on behalf of an association, business, labor union, etc.) You may 
review DOT's complete Privacy Act statement in the Federal Register 
published on April 11, 2000 (Volume 65, Number 70, Pages 19477-78), or 
you may visit http://dms.dot.gov.

List of Subjects in 49 CFR Part 563

    Incorporation by reference, Motor vehicle safety, Motor vehicles, 
Reporting and recordkeeping requirements.

0
For the reasons stated in the preamble, NHTSA hereby amends chapter V 
of title 49 of the Code of Federal Regulations by adding 49 CFR part 
563 to read as follows:

PART 563--EVENT DATA RECORDERS

Sec.
563.1 Scope.
563.2 Purpose.
563.3 Application.
563.4 Incorporation by reference.
563.5 Definitions.
563.6 Requirements for vehicles.
563.7 Data elements.
563.8 Data format.
563.9 Data capture.
563.10 Crash test performance and survivability.
563.11 Information in owner's manual.
563.12 Data retrieval tools.

    Authority: 49 U.S.C. 322, 30101, 30111, 30115, 30117, 30166, 
30168; delegation of authority at 49 CFR 1.50.


Sec.  563.1  Scope.

    This part specifies uniform, national requirements for vehicles 
equipped with event data recorders (EDRs) concerning the collection, 
storage, and retrievability of onboard motor vehicle crash event data. 
It also specifies requirements for vehicle manufacturers to make tools 
and/or methods commercially available so that crash investigators and 
researchers are able to retrieve data from EDRs.


Sec.  563.2  Purpose.

    The purpose of this part is to help ensure that EDRs record, in a 
readily usable manner, data valuable for effective crash investigations 
and for analysis of safety equipment performance (e.g., advanced 
restraint systems). These data will help provide a better understanding 
of the circumstances in which crashes and injuries occur and will lead 
to safer vehicle designs.


Sec.  563.3  Application.

    This part applies to the following vehicles manufactured on or 
after September 1, 2010, if they are equipped with an event data 
recorder: passenger cars, multipurpose passenger vehicles, trucks, and 
buses with a GVWR of 3,855 kg (8,500 pounds) or less and an unloaded 
vehicle weight of 2,495 kg (5,500 pounds) or less, except for walk-in 
van-type trucks or vehicles designed to be sold exclusively to the U.S. 
Postal Service. This part also applies to manufacturers of those 
vehicles. However, vehicles manufactured before September 1, 2011 that 
are manufactured in two or more stages or that are altered (within the 
meaning of 49 CFR 567.7) after having been previously certified to the 
Federal motor vehicle safety standards in accordance with Part 567 of 
this chapter need not meet the requirements of this part.


Sec.  563.4  Incorporation by reference.

    The materials listed in this section are incorporated by reference 
in the corresponding sections as noted. These incorporations by 
reference were approved by the Director of the Federal Register in 
accordance with 5 U.S.C. 522(a) and 1 CFR part 51. Copies of these 
materials may be inspected at the National Highway Traffic Safety 
Administration, Technical Information

[[Page 51044]]

Services, 400 Seventh Street, SW., Plaza Level, Room 403, Washington, 
DC 20590, or at the National Archives and Records Administration 
(NARA). For information on the availability of this material at NARA, 
call (202) 741-6030, or go to: http://www.archives.gov/federal_register/code_of_federal_regulations/ibr_locations.html.
    (a) The following materials are available for purchase from the 
Society of Automotive Engineers, Inc., 400 Commonwealth Drive, 
Warrendale, PA 15096-0001.
    (1) Society of Automotive Engineers (SAE) Recommended Practice 
J211-1 rev. March 1995, ``Instrumentation For Impact Test--Part 1--
Electronic Instrumentation'' SAE J211-1 (rev. March 1995) is 
incorporated by reference in Table 3 of Sec.  563.8;
    (2) [Reserved]
    (b) [Reserved]


Sec.  563.5  Definitions.

    (a) Motor vehicle safety standard definitions. Unless otherwise 
indicated, all terms that are used in this part and are defined in the 
Motor Vehicle Safety Standards, part 571 of this subchapter, are used 
as defined therein.
    (b) Other definitions.
    ABS activity means the anti-lock brake system (ABS) is actively 
controlling the vehicle's brakes.
    Air bag warning lamp status means whether the warning lamp required 
by FMVSS No. 208 is on or off.
    Capture means the process of buffering EDR data in a temporary, 
volatile storage medium where it is continuously updated at regular 
time intervals.
    Delta-V, lateral means the cumulative change in velocity, as 
recorded by the EDR of the vehicle, along the lateral axis, starting 
from crash time zero and ending at 0.25 seconds, and recorded every 
0.01 seconds.
    Delta-V, longitudinal means the cumulative change in velocity, as 
recorded by the EDR of the vehicle, along the longitudinal axis, 
starting from crash time zero and ending at 0.25 seconds, recorded 
every 0.01 seconds.
    Deployment time, frontal air bag means (for both driver and right 
front passenger) the elapsed time from crash time zero to the 
deployment command or for multi-staged air bag systems, the deployment 
command for the first stage.
    Disposal means the deployment command of the second (or higher, if 
present) stage of a frontal air bag for the purpose of disposing the 
propellant from the air bag device.
    End of event time means the moment at which the cumulative delta-V 
within a 20 ms time period becomes 0.8 km/h (0.5 mph) or less.
    Engine RPM means, for vehicles powered by internal combustion 
engines, the number of revolutions per minute of the main crankshaft of 
the vehicle's engine, and for vehicles not powered by internal 
combustion engines, the number of revolutions per minute of the motor 
shaft at the point at which it enters the vehicle transmission gearbox.
    Engine throttle, percent full means the driver requested 
acceleration as measured by the throttle position sensor on the 
accelerator pedal compared to the fully depressed position.
    Event means a crash or other physical occurrence that causes the 
trigger threshold to be met or exceeded.
    Event data recorder (EDR) means a device or function in a vehicle 
that records the vehicle's dynamic, time-series data during the time 
period just prior to a crash event (e.g., vehicle speed vs. time) or 
during a crash event (e.g., delta-V vs. time), intended for retrieval 
after the crash event. For the purposes of this definition, the event 
data do not include audio and video data.
    Frontal air bag means an inflatable restraint system that requires 
no action by vehicle occupants and is used to meet the applicable 
frontal crash protection requirements of FMVSS No. 208.
    Ignition cycle, crash means the number (count) of power cycles 
applied to the recording device at the time when the crash event 
occurred since the first use of the EDR.
    Ignition cycle download means the number (count) of power cycles 
applied to the recording device at the time when the data was 
downloaded since the first use of the EDR.
    Lateral acceleration means the component of the vector acceleration 
of a point in the vehicle in the y-direction. The lateral acceleration 
is positive from left to right, from the perspective of the driver when 
seated in the vehicle facing the direction of forward vehicle travel.
    Longitudinal acceleration means the component of the vector 
acceleration of a point in the vehicle in the x-direction. The 
longitudinal acceleration is positive in the direction of forward 
vehicle travel.
    Maximum delta-V, lateral means the maximum value of the cumulative 
change in velocity, as recorded by the EDR, of the vehicle along the 
lateral axis, starting from crash time zero and ending at 0.3 seconds.
    Maximum delta-V, longitudinal means the maximum value of the 
cumulative change in velocity, as recorded by the EDR, of the vehicle 
along the longitudinal axis, starting from crash time zero and ending 
at 0.3 seconds.
    Multi-event crash means the occurrence of 2 events, the first and 
last of which begin not more than 5 seconds apart.
    Non-volatile memory means the memory reserved for maintaining 
recorded EDR data in a semi-permanent fashion. Data recorded in non-
volatile memory is retained after a loss of power and can be retrieved 
with EDR data extraction tools and methods.
    Normal acceleration means the component of the vector acceleration 
of a point in the vehicle in the z-direction. The normal acceleration 
is positive in a downward direction and is zero when the accelerometer 
is at rest.
    Occupant position classification means the classification 
indicating that the seating posture of a front outboard occupant (both 
driver and right front passenger) is determined as being out-of-
position.
    Occupant size classification means, for right front passenger, the 
classification of an occupant as an adult and not a child, and for 
driver, the classification of the driver as not being of small stature.
    Pretensioner means a device that is activated by a vehicle's crash 
sensing system and removes slack from a vehicle safety belt system.
    Record means the process of saving captured EDR data into a non-
volatile device for subsequent retrieval.
    Safety belt status means the feedback from the safety system that 
is used to determine than an occupant's safety belt (for both driver 
and right front passenger) is fastened or not fastened.
    Seat track position switch, foremost, status means the status of 
the switch that is installed to detect whether the seat is moved to a 
forward position.
    Service brake, on and off means the status of the device that is 
installed in or connected to the brake pedal system to detect whether 
the pedal was pressed. The device can include the brake pedal switch or 
other driver-operated service brake control.
    Side air bag means any inflatable occupant restraint device that is 
mounted to the seat or side structure of the vehicle interior, and that 
is designed to deploy in a side impact crash to help mitigate occupant 
injury and/or ejection.
    Side curtain/tube air bag means any inflatable occupant restraint 
device that is mounted to the side structure of the vehicle interior, 
and that is designed to deploy in a side impact crash or rollover

[[Page 51045]]

and to help mitigate occupant injury and/or ejection.
    Speed, vehicle indicated means the vehicle speed indicated by a 
manufacturer-designated subsystem designed to indicate the vehicle's 
ground travel speed during vehicle operation.
    Stability control means any device that is not directly controlled 
by the operator (e.g., steering or brakes) and is intended to prevent 
loss of vehicle control by sensing, interpreting, and adjusting a 
vehicle's driving and handling characteristics, is controlling or 
assisting the driver in controlling the vehicle.
    Steering wheel angle means the angular displacement of the steering 
wheel measured from the straight-ahead position (position corresponding 
to zero average steer angle of a pair of steered wheels).
    Suppression switch status means the status of the switch indicating 
whether an air bag suppression system is on or off.
    Time from event 1 to 2 means the elapsed time from time zero of the 
first event to time zero of the second event.
    Time, maximum delta-V, longitudinal means the time from crash time 
zero to the point where the maximum value of the cumulative change in 
velocity is found, as recorded by the EDR, along the longitudinal axis.
    Time to deploy, pretensioner means the elapsed time from crash time 
zero to the deployment command for the safety belt pretensioner (for 
both driver and right front passenger).
    Time to deploy, side air bag/curtain means the elapsed time from 
crash time zero to the deployment command for a side air bag or a side 
curtain/tube air bag (for both driver and right front passenger).
    Time to first stage means the elapsed time between time zero and 
the time when the first stage of a frontal air bag is commanded to 
fire.
    Time to maximum delta-V, lateral means time from crash time zero to 
the point where the maximum value of the cumulative change in velocity 
is found, as recorded by the EDR, along the lateral axis.
    Time to nth stage means the elapsed time from the crash 
time zero to the deployment command for the nth stage of a frontal air 
bag (for both driver and right front passenger).
    Time zero means for systems with ``wake-up'' air bag control 
systems, the time occupant restraint control algorithm is activated; 
for continuously running algorithms, the first point in the interval 
where a longitudinal, cumulative delta-V of over 0.8 km/h (0.5 mph) is 
reached within a 20 ms time period; or for vehicles that record 
``delta-V, lateral,'' the first point in the interval where a lateral, 
cumulative delta-V of over 0.8 km/h (0.5 mph) is reached within a 5 ms 
time period.
    Trigger threshold means a change in vehicle velocity, in the 
longitudinal direction, that equals or exceeds 8 km/h within a 150 ms 
interval. For vehicles that record ``delta-V, lateral,'' trigger 
threshold means a change in vehicle velocity, in either the 
longitudinal or lateral direction that equals or exceeds 8 km/h within 
a 150 ms interval.
    Vehicle roll angle means the angle between the vehicle y-axis and 
the ground plane.
    Volatile memory means the memory reserved for buffering of captured 
EDR data. The memory is not capable of retaining data in a semi-
permanent fashion. Data captured in a volatile memory is continuously 
overwritten and is not retained in the event of a power loss or 
retrievable with EDR data extraction tools.
    X-direction means in the direction of the vehicle X-axis, which is 
parallel to the vehicle's longitudinal centerline. The X-direction is 
positive in the direction of forward vehicle travel.
    Y-direction means in the direction of the vehicle Y-axis, which is 
perpendicular to its X-axis and in the same horizontal plane as that 
axis. The Y-direction is positive from left to right, from the 
perspective of the driver when seated in the vehicle facing the 
direction of forward vehicle travel.
    Z-direction means in the direction of the vehicle Z-axis, which is 
perpendicular to the X- and Y-axes. The Z-direction is positive in a 
downward direction.


Sec.  563.6  Requirements for vehicles.

    Each vehicle equipped with an EDR must meet the requirements 
specified in Sec.  563.7 for data elements, Sec.  563.8 for data 
format, Sec.  563.9 for data capture, Sec.  563.10 for crash test 
performance and survivability, and Sec.  563.11 for information in 
owner's manual.


Sec.  563.7  Data elements.

    (a) Data elements required for all vehicles. Each vehicle equipped 
with an EDR must record all of the data elements listed in Table I, 
during the interval/time and at the sample rate specified in that 
table.

 Table I.--Data Elements Required for all Vehicles Equipped With an EDR
------------------------------------------------------------------------
                                     Recording interval/    Data sample
            Data element             time \1\  (relative   rate samples
                                        to time zero)       per second
------------------------------------------------------------------------
Delta-V, longitudinal..............  0 to 250 ms........             100
Maximum delta-V, longitudinal......  0-300 ms...........            N.A.
Time, maximum delta-V..............  0-300 ms...........            N.A.
Speed, vehicle indicated...........  -5.0 to 0 sec......               2
Engine throttle, % full (or          -5.0 to 0 sec......               2
 accelerator pedal, % full).
Service brake, on/off..............  -5.0 to 0 sec......               2
Ignition cycle, crash..............  -1.0 sec...........            N.A.
Ignition cycle, download...........  At time of download            N.A.
Safety belt status, driver.........  -1.0 sec...........            N.A.
Frontal air bag warning lamp, on/    -1.0 sec...........            N.A.
 off.
Frontal air bag deployment, time to  Event..............            N.A.
 deploy, in the case of a single
 stage air bag, or time to first
 stage deployment, in the case of a
 multi-stage air bag, driver.
Frontal air bag deployment, time to  Event..............            N.A.
 deploy, in the case of a single
 stage air bag, or time to first
 stage deployment, in the case of a
 multi-stage air bag, right front
 passenger.
Multi-event, number of events (1,2)  Event..............            N.A.
Time from event 1 to 2.............  As needed..........            N.A.
Complete file recorded (yes, no)...  Following other               N.A.
                                      data.
------------------------------------------------------------------------
\1\ Pre-crash data and crash data are asynchronous. The sample time
  accuracy requirement for pre-crash time is -0.1 to 1.0 sec (e.g., T =
  1 would need to occur between -1.1 and 0 seconds.)


[[Page 51046]]

    (b) Data elements required for vehicles under specified conditions. 
Each vehicle equipped with an EDR must record each of the data elements 
listed in column 1 of Table II for which the vehicle meets the 
condition specified in column 2 of that table, during the interval/time 
and at the sample rate specified in that table.

                    Table II.--Data Elements Required for Vehicles Under Specified Conditions
----------------------------------------------------------------------------------------------------------------
                                                                         Recording interval/time    Data sample
            Data element name              Condition for  requirement     1  (relative to time      rate  (per
                                                                                 zero)                second)
----------------------------------------------------------------------------------------------------------------
Lateral acceleration.....................  If recorded \2\...........  0-250 ms.................             500
Longitudinal acceleration................  If recorded...............  0-250 ms.................             500
Normal acceleration......................  If recorded...............  0-250 ms.................             500
Delta-V, lateral.........................  If recorded...............  0-250 ms.................             100
Maximum delta-V, lateral.................  If recorded...............  0-300 ms.................            N.A.
Time maximum delta-V, lateral............  If recorded...............  0-300 ms.................            N.A.
Time for maximum delta-V, resultant......  If recorded...............  0-300 ms.................            N.A.
Engine rpm...............................  If recorded...............  -5.0 to 0 sec............               2
Vehicle roll angle.......................  If recorded...............  -1.0 up to 5.0 sec 3.....              10
ABS activity (engaged, non-engaged)......  If recorded...............  -5.0 to 0 sec............               2
Stability control (on, off, engaged).....  If recorded...............  -5.0 to 0 sec............               2
Steering input...........................  If recorded...............  -5.0 to 0 sec............               2
Safety belt status, right front passenger  If recorded...............  -1.0 sec.................            N.A.
 (buckled, not buckled).
Frontal air bag suppression switch         If recorded...............  -1.0 sec.................            N.A.
 status, right front passenger (on, off,
 or auto).
Frontal air bag deployment, time to nth    If equipped with a          Event....................            N.A.
 stage, driver 4.                           driver's frontal air bag
                                            with a multi-stage
                                            inlator.
Frontal air bag deployment, time to nth    If equipped with a right    Event....................            N.A.
 stage, right front passenger 4.            front passenger's frontal
                                            air bag with a multi-
                                            stage inflator.
Frontal air bag deployment, nth stage      If recorded...............  Event....................            N.A.
 disposal, driver, Y/N (whether the nth
 stage deployment was for occupant
 restraint or propellant disposal
 purposes).
Frontal air bag deployment, nth stage      If recorded...............  Event....................            N.A.
 disposal, right front passenger, Y/N
 (whether the nth stage deployment was
 for occupant restraint or propellant
 disposal purposes).
Side air bag deployment, time to deploy,   If recorded...............  Event....................            N.A.
 driver.
Side air bag deployment, time to deploy,   If recorded...............  Event....................            N.A.
 right front passenger.
Side curtain/tube air bag deployment,      If recorded...............  Event....................            N.A.
 time to deploy, driver side.
Side curtain/tube air bag deployment,      If recorded...............  Event....................            N.A.
 time to deploy, right side.
Pretensioner deployment, time to fire,     If recorded...............  Event....................            N.A.
 driver.
Pretensioner deployment, time to fire,     If recorded...............  Event....................            N.A.
 right front passenger.
Seat track position switch, foremost,      If recorded...............  -1.0 sec.................            N.A.
 status, driver.
Seat track position switch, foremost,      If recorded...............  -1.0 sec.................            N.A.
 status, right front passenger.
Occupant size classification, driver.....  If recorded...............  -1.0 sec.................            N.A.
Occupant size classification, right front  If recorded...............  -1.0 sec.................            N.A.
 passenger.
Occupant position classification, driver.  If recorded...............  -1.0 sec.................            N.A.
Occupant position classification, right    If recorded...............  -1.0 sec.................           N.A.
 front passenger.
----------------------------------------------------------------------------------------------------------------
1 Pre-crash data and crash data are asynchronous. The sample time accuracy requirement for pre-crash time is -
  0.1 to 1.0 sec (e.g. T = -1 would need to occur between -1.1 and 0 seconds.)
2 ``If recorded'' means if the data is recorded in non-volatile memory for the purpose of subsequent
  downloading.
3 ``Vehicle roll angle'' may be recorded in any time duration, -1.0 sec to 5.0 sec is suggested.
4 List this element n-1 times, once for each stage of a multi-stage air bag system.

Sec.  563.8  Data format.

    (a) The data elements listed in Tables I and II, as applicable, 
must be recorded in accordance with the range, accuracy, resolution, 
and filter class specified in Table III.

                                    Table III.--Recorded Data Element Format
----------------------------------------------------------------------------------------------------------------
          Data element                   Range             Accuracy           Resolution         Filter class
----------------------------------------------------------------------------------------------------------------
Lateral acceleration............  -50 g to + 50 g...  5%....  0.01 g............  SAE J211-1,\1\
                                                                                               Class 60.
Longitudinal acceleration.......  -50 g to + 50 g...  5%....  0.01 g............  SAE J211-1,\1\
                                                                                               Class 60.
Normal Acceleration.............  -50 g to + 50 g...  5%....  0.01 g............  SAE J211-1,\1\
                                                                                               Class 60.
Longitudinal delta-V............  -100 km/h + 100 km/ 5%....  1 km/h............  N.A.
                                   h.
Lateral delta-V.................  -100 km/h to + 100  5%....  1 km/h............  N.A.
                                   km/h.
Maximum delta-V, longitudinal...  + 100 km/h + 100    5%....  1 km/h............  N.A.
                                   km/h.
Maximum delta-V, lateral........  -100 km/h to + 100  5%....  1 km/h............  N.A.
                                   km/h.

[[Page 51047]]

 
Time, maximum delta-V,            0-300 ms..........  3 ms..  2.5 ms............  N.A.
 longitudinal.
Time, maximum delta-V, lateral..  0-300 ms..........  3 ms..  2.5 ms............  N.A.
Time, maximum delta-V, resultant  0-300 ms..........  3 ms..  2.5 ms............  N.A.
Vehicle Roll Angle..............  -1080 deg to +      10 deg  10 deg............  N.A.
                                   1080 deg.
Speed, vehicle indicated........  0 km/h to 200 km/h  1 km/h  1 km/h............  N.A.
Engine throttle, percent full      0 to 100%........  5%....  1%................  N.A.
 (accelerator pedal percent
 full).
Engine rpm......................  0 to 10,000 rpm...  100     100 rpm...........  N.A.
                                                       rpm.
Service brake, on, off..........  On and Off........  N.A...............  On and Off........  N.A.
ABS activity....................  On and Off........  N.A...............  On and Off........  N.A.
Stability control (on, off,       On, Off, Engaged..  N.A...............  On, Off, Engaged..  N.A.
 engaged).
Steering wheel angle............  -250 deg CW to +    5 deg.  5 deg.............  N.A.
                                   250 deg CCW.
Ignition cycle, crash...........  0 to 60,000.......  1       1 cycle...........  N.A.
                                                       cycle.
Ignition cycle, download........  0 to 60,000.......  1       1 cycle...........  N.A.
                                                       cycle.
Safety belt status, driver......  On or Off.........  N.A...............  On or Off.........  N.A.
Safety belt status, right front   On or Off.........  N.A...............  On or Off.........  N.A.
 passenger.
Frontal air bag warning lamp      On or Off.........  N.A...............  On or Off.........  N.A.
 (on, off).
Frontal air bag suppression       On or Off.........  N.A...............  On or Off.........  N.A.
 switch status.
Frontal air bag deployment, time  0 to 250 ms.......  2 ms..  1 ms..............  N.A.
 to deploy/first stage, driver.
Frontal air bag deployment, time   0 to 250 ms......  2 ms..  1 ms..............  N.A.
 to deploy/first stage, right
 front passenger.
Frontal air bag deployment, time  0 to 250 ms.......  2 ms..  1 ms..............  N.A.
 to nth stage, driver.
Frontal air bag deployment, time   0 to 250 ms......  2 ms..  1 ms..............  N.A.
 to nth stage, right front
 passenger.
Frontal air bag deployment, nth   Yes/No............  N.A...............  Yes/No............  N.A.
 stage disposal, driver, y/n.
Frontal air bag deployment, nth   Yes/No............  N.A...............  Yes/No............  N.A.
 stage disposal, right front
 passenger, y/n.
Side air bag deployment, time to  0 to 250 ms.......  2 ms..  1 ms..............  N.A.
 deploy, driver.
Side air bag deployment, time to  0 to 250 ms.......  2 ms..  1 ms..............  N.A.
 deploy, right front passenger.
Side curtain/tube air bag         0 to 250 ms.......  2 ms..  1 ms..............  N.A.
 deployment, time to deploy,
 driver side.
Side curtain/tube air bag         0 to 250 ms.......  2 ms..  1 ms..............  N.A.
 deployment, time to deploy,
 right side.
Pretensioner deployment, time to  0 to 250 ms.......   2 ms.  1 ms..............  N.A.
 fire, driver.
Pretensioner deployment, time to  0 to 250 ms.......  2 ms..  1 ms..............  N.A.
 fire, right front passenger.
Seat track position switch,       Yes/No............  N.A...............  Yes/No............  N.A.
 foremost, status, driver.
Seat track position switch,       Yes/No............  N.A...............  Yes/No............  N.A.
 foremost, status, right front
 passenger.
Occupant size driver occupant     Yes/No............  N.A...............  Yes/No............  N.A.
 5th female size y/n.
Occupant size right front         Yes/No............  N.A...............  Yes/No............  N.A.
 passenger child y/n.
Occupant position                 Yes/No............  N.A...............  Yes/No............  N.A.
 classification, driver oop y/n.
Occupant position                 Yes/No............  N.A...............  Yes/No............  N.A.
 classification, right front
 passenger oop y/n.
Multi-event, number of events     1 or 2............  N.A...............  1 or 2............  N.A.
 (1, 2).
Time from event 1 to 2..........  0 to 5.0 sec......  0.1 sec...........  0.1 sec...........  N.A.
Complete file recorded (yes/no).  Yes/No............  N.A...............  Yes/No............  N.A.
----------------------------------------------------------------------------------------------------------------
\1\ Incorporated by reference, see Sec.   563.4.

    (b) Acceleration Time-History data and format: The longitudinal, 
lateral, and normal acceleration time-history data, as applicable, must 
be filtered in accordance with the filter class specified in Table III 
either during the recording phase or during the data downloading phase 
to include:
    (1) The Time Step (TS) that is the inverse of the sampling 
frequency of the acceleration data and which has units of seconds;
    (2) The number of the first point (NFP), which is an integer that 
when multiplied by the TS equals the time

[[Page 51048]]

relative to time zero of the first acceleration data point;
    (3) The number of the last point (NLP), which is an integer that 
when multiplied by the TS equals the time relative to time zero of the 
last acceleration data point; and
    (4) NLP-NFP+1 acceleration values sequentially beginning with the 
acceleration at time NFP*TS and continue sampling the acceleration at 
TS increments in time until the time NLP*TS is reached.


Sec.  563.9  Data capture.

    The EDR must capture and record the data elements for events in 
accordance with the following conditions and circumstances:
    (a) In an air bag deployment crash, the data recorded from any 
previous crash must be deleted (both events). The data related to the 
deployment must be captured and recorded. The memory must be locked to 
prevent any future overwriting of these data.
    (b) In an air bag non-deployment crash that meets the trigger 
threshold, delete all previously recorded data in the EDR's memory. 
Capture and record the current data, up to two events. In the case of 
two events, detection of the second event starts after the End of Event 
Time for event 1.


Sec.  563.10  Crash test performance and survivability.

    (a) Each vehicle subject to the requirements of S5, S14.5, S15, or 
S17 of 49 CFR 571.208, Occupant crash protection, must comply with the 
requirements in subpart (c) of this section when tested according to 
S8, S16, and S18 of 49 CFR 571.208.
    (b) Each vehicle subject to the requirements of 49 CFR 571.214, 
Side impact protection, that meets a trigger threshold or has a frontal 
air bag deployment, must comply with the requirements of subpart (c) of 
this section when tested according to the conditions specified in 49 
CFR 571.214 for a moving deformable barrier test.
    (c) The data elements required by Sec.  563.7, except for the 
``Engine throttle, percent full,'' ``engine RPM,'' and ``service brake, 
on/off,'' must be recorded in the format specified by Sec.  563.8, 
exist at the completion of the crash test, and be retrievable by the 
methodology specified by the vehicle manufacturer under Sec.  563.12 
for not less than 10 days after the test, and the complete data 
recorded element must read ``yes'' after the test.


Sec.  563.11  Information in owner's manual.

    (a) The owner's manual in each vehicle covered under this 
regulation must provide the following statement in English:

    This vehicle is equipped with an event data recorder (EDR). The 
main purpose of an EDR is to record, in certain crash or near crash-
like situations, such as an air bag deployment or hitting a road 
obstacle, data that will assist in understanding how a vehicle's 
systems performed. The EDR is designed to record data related to 
vehicle dynamics and safety systems for a short period of time, 
typically 30 seconds or less. The EDR in this vehicle is designed to 
record such data as:
     How various systems in your vehicle were operating;
     Whether or not the driver and passenger safety belts 
were buckled/fastened;
     How far (if at all) the driver was depressing the 
accelerator and/or brake pedal; and,
     How fast the vehicle was traveling.
    These data can help provide a better understanding of the 
circumstances in which crashes and injuries occur. NOTE: EDR data 
are recorded by your vehicle only if a non-trivial crash situation 
occurs; no data are recorded by the EDR under normal driving 
conditions and no personal data (e.g., name, gender, age, and crash 
location) are recorded. However, other parties, such as law 
enforcement, could combine the EDR data with the type of personally 
identifying data routinely acquired during a crash investigation.
    To read data recorded by an EDR, special equipment is required, 
and access to the vehicle or the EDR is needed. In addition to the 
vehicle manufacturer, other parties, such as law enforcement, that 
have the special equipment, can read the information if they have 
access to the vehicle or the EDR.

    (b) The owner's manual may include additional information about the 
form, function, and capabilities of the EDR, in supplement to the 
required statement in Sec.  563.11(a).


Sec.  563.12  Data retrieval tools.

    Each manufacturer of a motor vehicle equipped with an EDR shall 
ensure by licensing agreement or other means that a tool(s) is 
commercially available that is capable of accessing and retrieving the 
data stored in the EDR that are required by this part. The tool(s) 
shall be commercially available not later than 90 days after the first 
sale of the motor vehicle for purposes other than resale.

    Issued on: August 18, 2006.
Nicole R. Nason,
Administrator.
[FR Doc. 06-7094 Filed 8-21-06; 10:00 am]
BILLING CODE 4910-59-P